Dickimaw Books - Gallery (All Predefined Glossary Styles)

In general, it’s best to prevent unnecessary packages from being loaded, so I recommend you use nolong and nosuper, at the very least, if you don’t require the “long” or “super” styles.

The following files containing test entries are used in the examples: example-glossaries-brief.tex, example-glossaries-symbols.tex, example-glossaries-childnoname.tex, example-glossaries-childmultipar.tex, example-glossaries-parent.tex, example-glossaries-multipar.tex or example-glossaries-user.tex. These are all provided with the glossaries package. If you have the glossaries package installed, they should already be on your path. If you get any undefined control sequence or undefined style errors, then you probably need to update your version of glossaries (or glossaries-extra, where appropriate). You can click on the sample images below to view the PDF.

The bookindex, longextra, topic and table styles require the glossaries-extra extension package. The bib2gls examples in those sections use entries defined in sample .bib files provided by the glossaries-extra package and the testidx-glossaries package.

The number 1 appearing in red after the description indicates the position of the location (number) list. The locations are all page 1 in these examples, as the entries were all indexed on the first page of the sample document. In general, the location list will typically contain a comma-separated list of page numbers or number ranges. The number list can be suppressed using the nonumberlist option. (The number appears in red because the hyperref package has been loaded with internal hyperlinks set to that colour.)

Package options that affect the appearance of predefined styles are listed in section 2.3 Glossary Appearance Options of the glossaries user manual.

The glossary is typically divided into groups: symbols, numbers and then each letter group. In the case of the letter groups, the identifying group is given by the initial letter of the sort key. Most of the styles provide a visual separation between the letter groups. This may be either just a vertical gap or a group heading. If your glossary hasn’t been sorted alphabetically, then don’t have a visual indicator between groups, as the grouping no longer makes sense. Most of the basic predefined styles allow you to suppress the vertical gap between groups with the nogroupskip option.

Note that some styles use the post description hook \glspostdescription. This displays a full stop and adjusts the space factor if nopostdot=false and does nothing if nopostdot=true. The hook can be suppressed on an individual basis by using \nopostdesc in the entry’s description. The glossaries-extra package modifies \glspostdescription so that it first does \glsxtrpostdescription, which allows further customization according to the entry’s category. Version 1.02 of glossaries-extra provides patches to all the predefined styles that are missing \glspostdescription through the glossaries-extra-stylemods package. For example, to patch the styles in the glossary-longbooktabs package:

\usepackage{glossaries-extra}
\usepackage{glossary-longbooktabs}
\usepackage{glossaries-extra-stylemods}

\usepackage[stylemods=longbooktabs]{glossaries-extra}

For other examples of glossary styles, see the main gallery.

glossary-tree

tree* 🔗

The tree* style is new to glossaries version 4.59 (but make sure you have the corrected v4.6) and is more flexible than the older styles provided by the glossary-tree package, listed below. Unlike the older styles, which have user-level commands that may be redefined to adjust the style, the tree* is mostly adjusted via a key=value interface. For example:

\GlsTreeSetup{
 group-headings,
 name-style={name symbol}
}

\setupglossaries{
 style-options =
  {
    tree* =
     {
       group-headings,
       name-style={name symbol}
     }
  }
}

The test files for tree* use \makenoidxglossaries and \printnoidxglossary which means they only require two LaTeX calls to create the completed document. However, you will not only need to ensure that you have glossaries v4.6+ installed, but will also need datatool v3.0+.

Note that datatool v3.0+ has localisation support, which needs to be installed separately. At the time of writing, there is only limited support. Only the datatool language support (not the region support) is applicable to \printnoidxglossary (see datatool-english). The test document doesn't specify any language support, so sorting is mostly by character code (first converting to lowercase to ensure a case-insensitive order).

tree* style: default settings 🔗

The tree* style is designed for hierarchical glossaries, which can be demonstrated with the example-glossaries-parent.tex file:

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage[style=tree*]{glossaries}
\makenoidxglossaries
\loadglsentries{example-glossaries-parent}
\glsaddall
\begin{document}
\printnoidxglossary
\end{document}

The thumbnail image below is truncated at the end of the “N” letter group. Click on the image to view the PDF of the complete document.

The names are shown in bold and each hierarchical level is indented. This set of test entries don’t have symbols, but if they did, the symbol would appear in parentheses after the name and before the description.

This looks very similar to the older tree style. The default gap between letter groups (the “group skip”) is slightly shorter with the tree* than with the tree style.

tree* style: letter group headings 🔗

The group-headings boolean option will show header groups:

\GlsTreeSetup{group-headings}

This now has letter groups headings (A, C and N), flush left and in bold. This looks very similar to the treegroup style, but the vertical gap between the heading and the entry following it is smaller with the tree* style.

tree* style: sub-groups 🔗

The tree* style also supports sub-groups, which is only available with bib2gls and the group-level resource option. This example will require glossaries-extra (and glossaries v4.6):

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage[style=tree*,nostyles,stylemods={tree},record]{glossaries-extra}
\BibGlsOptions{group}
\GlsXtrLoadResources[
 src={example-glossaries-parent},
 selection=all,
 group-level=all
]
\GlsTreeSetup{
 group-headings
}
\begin{document}
\printunsrtglossary
\end{document}

tree* style: hyper-navigation strip 🔗

The hyper-nav option can be combined with the group-headings option from the earlier groups example to add a navigation strip at the start of the glossary with hyperlinks to each group.

\GlsTreeSetup{
 group-headings,
 hyper-nav
}

This requires an extra LaTeX run to get the document up-to-date.

The thumbnail image below is again truncated at the end of the “N” letter group, but the navigation strip shows all the letter groups included in the glossary.

tree* style: names and symbols 🔗

You may instead prefer PDF bookmarks to help navigate around a large glossary. This is best done with the toc option:

\setupglossaries{
 toc,
 style-options = 
  {
    tree* =
     {
       group-headings,
       bookmark-groups
     }
  }
}

Again, this requires an extra LaTeX run to get the document up-to-date.

The glossary is as for the earlier example with the group-headings setting on, but now there are PDF bookmarks.

The exact style of the bookmarks varies according to the PDF viewer, but most have a tree-like format with collapsible elements.

tree* style: names and symbols 🔗

The example-glossaries-parent.tex file used in the above examples doesn’t have any symbols. The example-glossaries-user.tex file, new to glossaries v4.59, has hierarchical entries with symbols. The name key is set to the notation representing a mathematical constant, such as \(\delta_S\) for the silver ratio, and the symbol key is set to the formula, such as \(\sqrt{2}+1\). In some cases, such as for pi \(\pi\), the symbol key isn’t set.

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage[style=tree*]{glossaries}
\makenoidxglossaries
\loadglsentries{example-glossaries-constants}
\glsaddall
\begin{document}
\printnoidxglossary
\end{document}

The default style governing the name and symbol is name-style={name (symbol)}, which means that the name comes first, followed by (if set) the symbol in parentheses. This looks a bit odd for the psi \(\psi\) supergolden ratio, which has quite a tall symbol.

tree* style: fixed width name and symbol 🔗

The default setting is for the name and symbol to simply take up their natural width, but it’s possible to put them in fixed width boxes. You can either set separate widths for the name and symbol or a combined width. You can specify an explicit dimension, such as name-width=2.5cm, or you can specify the widest name or symbol and indicate that the corresponding box should be the width of that name or symbol.

This can be done by first typesetting the glossary with the default natural width setting and then inspecting the PDF to visually determine which name and symbol is the widest. Alternatively, if you are using \printnoidxglossary or \printunsrtglossary, you can make use of the applicable loop item hook to perform the calculations.

With glossaries v4.59+, the applicable hook with \printnoidxglossary is \glsnoidxitemhook:

\renewcommand{\glsnoidxitemhook}[2]{%
 \GlsTreeUpdateWidestNameOrSymbol[#1]{#2}%
}

The applicable hook with \printunsrtglossary is \printunsrtglossaryentryprocesshook:

\renewcommand{\printunsrtglossaryentryprocesshook}[1]{%
 \GlsTreeUpdateWidestNameOrSymbol{#1}%
}

There is no convenient hook for use with \printglossary, as the code for typesetting the glossary is simply input from a file that’s created by makeindex or xindy. You could iterate over all entries with \forglsentries before \printglossary but that may include entries that have been defined but not indexed, which could affect the calculations. The other possibility is to use the pre-item hook \glossaries_tree_pre_item:nnn to make the calculations while the glossary is being typeset, save the result in the aux file and fetch it on the next run, but that complicates the document build.

The width settings should now be set to “widest”. The parentheses around the symbol can be removed by changing the name style:

\GlsTreeSetup{
  name-width=widest,
  symbol-width=widest,
  sub-name-width=widest,
  sub-symbol-width=widest,
  name-style={name symbol}
}

tree* style: aligning name, symbol and short descriptions 🔗

Note that although the name and symbol are set in separate fixed width boxes, which aligns them according to the hierarchical level, the description still follows on normally from the name and (if set) symbol in the same paragraph. Since some entries don’t have a symbol, this means that the descriptions aren’t aligned. The description wraps around to the next line according to the normal paragraph behaviour.

You can additionally set a fixed width for the outer box that encompasses the name and symbol, and calculate the hanging indentation required to align the descriptions:

\GlsTreeSetup{
  name-width=widest,
  symbol-width=widest,
  sub-name-width=widest,
  sub-symbol-width=widest,
  name-style={name symbol},
  name-symbol-width=widest,
  sub-name-symbol-width=widest,
  hang-indent=calculated
}

tree* style: aligning short descriptions 🔗

If you don’t need to have the symbols aligned, but you still want the descriptions aligned, then you need a fixed width for the outer box but not for the inner name and symbol boxes. You still need to supply the widest name and symbol if you want the dimension to be calculated, but now it needs to the name and symbol for the entry where the combined width is wider than for any other entry for the same hierarchical level.

\renewcommand{\glsnoidxitemhook}[2]{%
 \GlsTreeUpdateWidestNameAndSymbol[#1]{#2}%
}

\GlsTreeSetup{
  name-symbol-width=widest,
  sub-name-symbol-width=widest,
  name-style={name symbol},
  hang-indent=calculated
}

Note that this will only align the top-level descriptions. The sub-levels will have a fixed width box for the name and symbol area but the hanging indentation for sub-levels will prevent them from being aligned if the description is long enough to line wrap. So this technique is only really designed for short child paragraphs or no child entries.

tree* style: omitting the name and symbol for child entries 🔗

Other variations include omitting the symbol (name-style=name) or swapping the name and symbol round (name-style={symbol name}) or omitting the child name and symbol. For example, the following omits the symbol for top-level entries and omits both the name and symbol for child entries. The subentrycounter option is used to number level 1 entries:

\usepackage[style=tree*,subentrycounter]{glossaries}
\makenoidxglossaries
\loadglsentries{example-glossaries-constants}
\glsaddall

\GlsTreeSetup{
 name-style=name,
 child-name-style=omit
}

tree* style: multi-paragraph descriptions 🔗

The example-glossaries-childmultipar.tex file contains top-level and level 1 entries with long descriptions. The test document is now just:

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage[style=tree*]{glossaries}
\makenoidxglossaries
\loadglsentries{example-glossaries-childmultipar}
\glsaddall
\begin{document}
\printnoidxglossary
\end{document}

tree* style: setting hang-indent and par-indent 🔗

The tree* style is the only style provided by glossary-tree.sty that works well with multi-paragraph descriptions with its default settings.

If you change the hanging indentation, you will likely also need to change the paragraph indentation. For example:

\GlsTreeSetup{
  hang-indent=2em,
  par-indent=3em
}

tree* style: calculated hang-indent 🔗

The example-glossaries-childmultipar.tex file doesn’t have any descriptions, so if you want a fixed width box, you only need to calculate the widest name. If you are using bib2gls and the name only contains text, then you can use the set-widest resource option. With \printnoidxglossary, you can use:

\ExplSyntaxOn
\renewcommand{\glsnoidxitemhook}[2]{
  \glossaries_tree_update_widest_name:nn
   { #1 }
   { \glsentryname { #2 } }
}
\ExplSyntaxOff

\renewcommand{\glsnoidxitemhook}[2]{
 \ifcase#1 
   \GlsTreeSetup{update-widest-name=\glsentryname{#2}}%
 \or
   \GlsTreeSetup{update-widest-sub-name=\glsentryname{#2}}%
 \or
   \GlsTreeSetup{update-widest-sub-sub-name=\glsentryname{#2}}%
 \fi
}

\GlsTreeSetup{
  name-style=name,
  name-symbol-width=widest,
  hang-indent=calculated,
  par-indent={\hangindent+1em}
}

tree* style: paragraph break before description 🔗

If you want the description to start a new paragraph, remember that the key=value interface doesn’t allow \par, but you can instead use \glspar. The following starts a new paragraph and suppresses the paragraph indentation at the start of the description, but a horizontal space the width of the hanging indentation is need to align the first line of the paragraph with the remainder.

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage[style=tree*]{glossaries}
\makenoidxglossaries
\loadglsentries{example-glossaries-childmultipar}
\glsaddall
\GlsTreeSetup{
  hang-indent=2em,
  par-indent=3em,
  pre-description={\glspar\noindent\hspace{\hangindent}}
}
\begin{document}
\printnoidxglossary
\end{document}

index 🔗

This has a similar look to the list style but it's more compact, without the excess spacing caused by the description environment.

To illustrate the way this style displays symbols, the test file can be modified to use example-glossaries-symbols.tex instead of example-glossaries-brief.tex. This will also require the package option nopostdot as the symbol dummy entries contain the required sentence terminators.

To illustrate the way this style displays child entries, the test file can be modified to use example-glossaries-parent.tex.

The child names are shown in bold, just like the top-level names, but each hierarchical level is indented. This set of test entries don’t have symbols, but if they did, the child symbol would appear in parentheses before the description, just like the top-level symbols.

The left indentation for the top-level entries is created with \glstreeitem, which just uses \@idxitem (if defined) by default, which is used by standard indexes. The indentation is actually a hanging indent, created with \hangindent, rather than a paragraph indent. This means that the name is flush with the left margin, but subsequent lines in the same paragraph will all be indented.

Level 1 sub-entries start with \glstreesubitem, which behaves like \subitem, used in standard indexes. This does the same hanging indent as \glstreeitem, followed by a horizontal space of 20pt. This means that if a child entry has a long description, then subsequent lines of the description will simply have the same level of indentation as the top-level.

Level 2 sub-entries start with a similar command \glstreesubsubitem. Any deeper levels will be treated as a level 2.

Note that the index style isn’t well-suited for multi-paragraph descriptions. To illustrate this, the test file can be modified to use example-glossaries-multipar.tex.

Only the first paragraph of the description for each entry obeys the hanging indent. The subsequent paragraphs have lost the left indent.

This can be overcome for multi-paragraph descriptions in top-level entries with glossaries v4.26 and the following adjustment:

\renewcommand{\glstreeitem}{%
 \parindent0pt\par\hangindent40pt
 \everypar{\parindent50pt\hangindent40pt}}

Now the second paragraph has the left indent and has the additional paragraph indentation for the first line of the paragraph. (Child entries will need to have \glstreesubitem and \glstreesubsubitem redefined, as appropriate.)

This provides a convenient alternative to the list style. An alternative to altlist can be achieved by adding:

\renewcommand{\glstreepredesc}{\par
  \glstreeitem\parindent\hangindent}

This has the name in bold (flush with the left margin), and the first paragraph starting with a hanging indent on the next line. Both paragraphs are indented as before.

A better method with glossaries v4.6+ is to use the tree* style instead.

indexgroup 🔗

The group heading is followed by \indexspace, which is also what \glsgroupskip does unless the nogroupskip setting is used. This means that with the default setting, there will be an equal vertical space above and below the group headings (except for the first one), but if you use \glsgroupskip you will have no space before the heading, but still have space after it.

With the glossaries-extra-stylemods package, the vertical space is added with \glstreegroupheaderskip which prevents a page break.

indexhypergroup 🔗

tree 🔗

With no sub-entries or long lines, this appears the same as the index style. Unlike the index style, the tree style sets both the paragraph indent and hanging indent to 0pt.

To illustrate the way this style displays child entries, the test file can be modified to use example-glossaries-parent.tex.

With no long lines, this appears the same as the index style.

Sub entries have the paragraph indent and hanging indent set to n times the value of \glstreeindent (10pt by default), where n is the hierarchical level.

treegroup 🔗

With no long lines, this appears the same as the indexgroup style. As with the indexgroup style, \indexspace is placed after the group title. With the glossaries-extra-stylemods package, the vertical space is added with \glstreegroupheaderskip which prevents a page break.

treehypergroup 🔗

With no long lines, this appears the same as the indexhypergroup style.

treenoname 🔗

treenonamegroup 🔗

treenonamehypergroup 🔗

alttree 🔗

\glssetwidest[level]{widest name}

\glssetwidest{consectetuer}

The drawback with this is that it means you need to know the widest name or work it out programmatically by looping over all entries. As from v4.22, the glossary-tree package provides

\glsfindwidesttoplevelname[glossary list]

\glsfindwidesttoplevelname
\setglossarystyle{alttree}
\printglossaries

\renewcommand{\glossarypreamble}{\glsfindwidesttoplevelname[\currentglossary]}
\setglossarystyle{alttree}
\printglossaries

As with the tree style, a vertical gap is inserted between letter groups, so use the nogroupskip option if your entries aren’t sorted alphabetically.

This has the appearance of a two column table with the name in the first “column”, without actually using a tabular-like environment. The effect is achieved by setting the hanging indent and paragraph indent to the width of the widest name and then starting the paragraph with a right-justified zero-width box that contains a left-justified box of the indent width, which contains the name. This shifts the name leftwards into the indent space, allowing the remaining content (symbol, if present, description and number list) to be formatted in the indented paragraph, which can be broken across a page (unlike a paragraph cell in a longtable environment). As with the index and tree styles, the hanging indent is reset at the end of the paragraph.

The location of the zero-width box is shown in red below:

If a zero-width box contains left-justified text, the text will overlap the content to the right.

If a zero-width box contains right-justified text, the text will overlap the content to the left. In this case, the content on the left is the empty space caused by the indent.

This would make the name right-justified, which is why a left-justified inner box of the same width as the indent is needed.

If the widest name for any sub-entry level hasn’t been specified, the child names are given the same width box as the top-level names. This can lead to strange results. In the sample below I’ve used the example-glossaries-parent.tex file with:

\renewcommand{\glossarypreamble}{%
 \glsfindwidesttoplevelname[\currentglossary]%
}

This can be corrected with:

\glssetwidest[1]{nonummy quis}
\glssetwidest[2]{duisnibh}

This gives a tabbing effect, where each sub-level has been tabbed further to the right, without actually using the tabbing environment.

alttreegroup 🔗

As with the treegroup, each heading is followed by \indexspace, regardless of the nogroupskip setting (which only affects the space before the heading).

alttreehypergroup 🔗

The navigation line is as the treehypergroup style.

glossary-mcols

\usepackage{glossary-mcols}

Except for the new mcoltree* style, the number of columns is given by \glsmcols, which defaults to 2. For example, to have a three column glossary:

\renewcommand{\glsmcols}{3}
\setglossarystyle{mcolindex}

These styles essentially “inherit” from the corresponding glossary-tree style, so they use the same formatting commands. The adjustments are mostly limited to redefining the theglossary environment.

Take care not to confuse this with the columns within the tabular-like styles, such as the long style.

If you don’t want the description displayed, you may prefer the bookindex style provided with the glossaries-extra extension package. Alternatively use the mcoltree* style with the tree* boolean omit-description option.

mcoltree* 🔗

The mcoltree* style is new to glossaries version 4.59 (but make sure you have the corrected v4.6) and is more flexible than the older styles provided by the glossary-mcols package, listed below. It’s based on the tree* style, and essentially just wraps the tree* style glossary code inside either the multicols or multicols* environment. It will therefore reflect the current tree* settings. Options that are specific to mcoltree* need to be set with the mcoltree* key within style-options. For example:

\setupglossaries{
 style-options =
  {
    mcoltree* = 
     {
        columns=3
     },
    tree* =
     {
       group-headings,
       name-style={name symbol}
     }
  }
}

\documentclass{article}
\usepackage[colorlinks]{hyperref}
\usepackage{glossaries}
\usepackage{glossary-mcols}
\makenoidxglossaries
\loadglsentries{example-glossaries-brief}
\glsaddall
\begin{document}
\printnoidxglossary[style=mcoltree*]
\end{document}

Letter group headings and PDF bookmarks can be added as for the earlier tree* example. The descriptions may be omitted and the bold font removed from the name. The alignment for the letter group headings can also be adjusted. The following settings produce a similar result to the bookindex style.

\setupglossaries{
 toc,
 style-options = 
  {
    tree* =
     {
       group-headings,
       bookmark-groups
       name-font={},
       omit-description,
       header-align=center,
     }
  }
}

mcolindex 🔗

The effect is similar to using the index style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcolindexgroup 🔗

The effect is similar to using the indexgroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcolindexhypergroup 🔗

The effect is similar to using the indexhypergroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that this style bunches up the navigation line at the start of the first column (which is what you get if you put the indexhypergroup style in a multi-column page). If you feel this is inappropriate, the next style mcolindexspannav may look better.

mcolindexspannav 🔗

The effect is like the mcolindexgroup style with the navigation line spanning the line width below the section header.

mcoltree 🔗

The effect is similar to using the tree style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcoltreegroup 🔗

The effect is similar to using the treegroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcoltreehypergroup 🔗

The effect is similar to using the treehypergroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that this style bunches up the navigation line at the start of the first column. If you feel this is inappropriate, the next style mcoltreespannav may look better.

mcoltreespannav 🔗

The effect is like the mcoltreegroup style with the navigation line spanning the line width below the section header.

mcoltreenoname 🔗

The effect is similar to using the treenoname style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcoltreenonamegroup 🔗

The effect is similar to using the treenonamegroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

mcoltreenonamehypergroup 🔗

The effect is similar to using the treenonamehypergroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that this style bunches up the navigation line at the start of the first column (which isn’t particularly obvious in this example). If you feel this is inappropriate, the next style mcoltreenonamespannav may look better.

mcoltreenonamespannav 🔗

The effect is like the mcoltreenonamegroup style with the navigation line spanning the line width below the section header.

mcolalttree 🔗

The effect is similar to using the alttree style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that the columns are slightly unbalanced at the top, with the first line of the first column lower than the first line of the second column. If you are using the glossaries-extra package, this effect is corrected by the glossaries-extra-stylemods package, which shifts the initialisation code outside of the multicols environment.

mcolalttreegroup 🔗

The effect is similar to using the alttreegroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that, as with the mcolalttree style, the columns are slightly unbalanced at the top. Since the mcolalttreegroup style uses the mcolalttree style, the correction applied by the glossaries-extra-stylemods package, will also correct this style. Likewise for the following styles.

mcolalttreehypergroup 🔗

The effect is similar to using the alttreehypergroup style in a two-column page, except that the multicols environment balances the columns to make them approximately the same length.

Note that this style bunches up the navigation line at the start of the first column. If you feel this is inappropriate, the next style mcolalttreespannav may look better.

mcolalttreespannav 🔗

The effect is like the mcolalttreegroup style with the navigation line spanning the line width below the section header.

glossary-inline

\usepackage{glossary-inline}

inline 🔗

 \renewcommand*{\glossarysection}[2][]{} 

 \renewcommand*{\glossarysection}[2][]{\textbf{#2}: } 

The name is formatted using \glsinlinenameformat. The default definition is:

\newcommand*{\glsinlinenameformat}[2]{\glstarget{#1}{#2}}

\renewcommand*{\glsinlinenameformat}[2]{\glstarget{#1}{\textsc{#2}}}

One-level sub-entries are supported, but the child name won’t be displayed. For example, using the example-glossaries-childnoname.tex dummy entries and the subentrycounter option, which automatically numbers the sub-entries:

Note that this inserts a colon between the top-level description and the sub-level entry.

glossary-list

All the list styles ignore the symbol field. You may prefer one of the tree-like styles instead.

Most of the list styles are incompatible with classicthesis. If you are using the classicthesis package with glossaries, you will need to change the default style (unless you are using glossaries v4.26+ which checks for classicthesis and switches the default style to index).

list 🔗

Note that the list style doesn’t display the names of sub-entries, and generally doesn’t work well with child entries. If I change the test file to example-glossaries-childnoname.tex then the glossary is as shown below. The space before the location list and the dot after it are hard-coded into the sub-entry style, so this style isn’t suitable for sub-entries with the location lists suppressed. If you have sub-entries, consider one of the other styles instead. (Versions below 4.22 were missing the space before the sub-entry descriptions, which made the location lists run into the next description, but even with that fixed, it’s still not a good style for glossaries with child entries.)

listgroup 🔗

listhypergroup 🔗

altlist 🔗

The altlist style allows one-level child entries and will start a new paragraph for each child entry, but the child name won’t be displayed. In the sample below I’ve used the example-glossaries-childnoname.tex file with the subentrycounter package option, which automatically numbers the sub-entries. This just makes them a bit easier to see in the sample and distinguishes them from the parent entry’s description.

altlistgroup 🔗

altlisthypergroup 🔗

listdotted 🔗

This style is inappropriate for child entries. In the sample below, I’ve used the example-glossaries-parent.tex file with the subentrycounter option on, which automatically numbers the child entries.

The numbering helps to highlight the sub-entries but the hierarchy has been flattened to just the top-level and one sub-level.

sublistdotted 🔗

The above shows how poorly this style performs for glossaries with only top-level entries, since all it does is show the name in bold. If the sample file switches from example-glossaries-brief.tex to example-glossaries-parent.tex for the dummy entries, the result is much better:

glossary-long

Similar styles are available with the glossary-longragged, glossary-longbooktabs, and glossary-longextra packages.

Take care not to confuse these table styles with the mcols styles.

long 🔗

\begin{longtable}{lp{\glsdescwidth}}

\setglossarystyle{long}
\renewenvironment{theglossary}%
 {\begin{longtable}[l]{lp{\glsdescwidth}}}%
 {\end{longtable}}

An extra vertical gap is inserted between letter groups, so use the nogroupskip option if your entries haven’t been sorted alphabetically.

This style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries.

longborder 🔗

longheader 🔗

The header text for the first column is given by \entryname (Notation), and the header text for the second column is given by \descriptionname (Description). The header text is formatted in bold.

longheaderborder 🔗

The header row is as for the longheader style.

The border lines are as for the longborder style, with an extra horizontal line after the header row.

long3col 🔗

\begin{longtable}{lp{\glsdescwidth}p{\glspagelistwidth}}

As with the long style, this style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries.

long3colborder 🔗

long3colheader 🔗

The header text for the first column is given by \entryname (Notation), the header text for the second column is given by \descriptionname (Description), and the header text for the third column is given by \pagelistname (Page List). The header text is formatted in bold.

long3colheaderborder 🔗

long4col 🔗

\begin{longtable}{llll}

Note that the test file loads example-glossaries-brief.tex, which doesn’t provide symbols for any of the dummy entries, so the symbol column is blank in the sample below. (The example-glossaries-symbols.tex test file has descriptions that are too long for this style.)

As with the long style, this style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries.

If you have glossaries v4.51+, there’s an extra test file example-glossaries-user.tex which has both symbols and short descriptions, and also has child entries, and so is a better set to demonstrate this style (the subentrycounter hasn’t been used):

long4colborder 🔗

Compare with the example-glossaries-user.tex test file, which has symbols for top level entries and most (but not all) child entries:

long4colheader 🔗

Compare with the example-glossaries-user.tex test file, which has symbols for top level entries and most (but not all) child entries:

The left-alignment of the symbol column is more obvious in this case.

long4colheaderborder 🔗

altlong4col 🔗

\begin{longtable}{lp{\glsdescwidth}lp{\glspagelistwidth}}

To illustrate the way this style displays symbols, the test file can be modified to use example-glossaries-symbols.tex instead of example-glossaries-brief.tex. A vertical gap is inserted between letter groups, so use the nogroupskip option if your entries haven’t been sorted alphabetically. This style doesn’t use the post description hook (unless you use glossaries-extra-stylemods).

As with the long style, this style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries. Since this set of test entries don’t have symbols, the symbol column is empty.

Compare with the example-glossaries-user.tex test file, which has symbols for top level entries and most (but not all) child entries:

altlong4colborder 🔗

altlong4colheader 🔗

altlong4colheaderborder 🔗

glossary-longragged

\usepackage{glossary-longragged}

\usepackage{glossary-longbooktabs}

longragged 🔗

\begin{longtable}{l>{\raggedright}p{\glsdescwidth}}

The result is like the long style but has ragged right formatting in the description column.

longraggedborder 🔗

The result is like the longborder style but has ragged right formatting in the description column.

longraggedheader 🔗

The result is like the longheader style but has ragged right formatting in the description column.

longraggedheaderborder 🔗

The result is like the longheaderborder style but has ragged right formatting in the description column.

longragged3col 🔗

\begin{longtable}{l>{\raggedright}p{\glsdescwidth}%
       >{\raggedright}p{\glspagelistwidth}}

The result is like the long3col style but has ragged right formatting in the description and page list columns.

longragged3colborder 🔗

The result is like the long3colborder style but has ragged right formatting in the description and page list columns.

longragged3colheader 🔗

The result is like the long3colheader style but has ragged right formatting in the description and page list columns.

longragged3colheaderborder 🔗

The result is like the long3colheaderborder style but has ragged right formatting in the description and page list columns.

altlongragged4col 🔗

\begin{longtable}{l>{\raggedright}p{\glsdescwidth}l%
       >{\raggedright}p{\glspagelistwidth}}

(Note that there is no analogous ragged style to the long4col style, as that style doesn’t have any paragraph columns.)

As with the altlong4col style, an extra vertical gap is inserted between letter groups, so use the nogroupskip option if your entries haven’t been sorted alphabetically. Note that this style doesn’t use the post description hook (unless you use glossaries-extra-stylemods), so the nopostdot setting has no effect.

The result is like the altlong4col style but has ragged right formatting in the description and page list columns.

altlongragged4colborder 🔗

The result is like the altlong4colborder style but has ragged right formatting in the description and page list columns.

altlongragged4colheader 🔗

The result is like the altlong4colheader style but has ragged right formatting in the description and page list columns.

altlongragged4colheaderborder 🔗

The result is like the altlong4colheaderborder style but has ragged right formatting in the description and page list columns.

glossary-longbooktabs

\usepackage{glossary-longbooktabs}

If you are using glossaries-extra.sty, the longextra styles can be used to provide more flexibility.

long-booktabs 🔗

The result is like the longheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the longheaderborder style there are no vertical rules.

long3col-booktabs 🔗

The result is like the long3colheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the long3colheaderborder style there are no vertical rules.

long4col-booktabs 🔗

The result is like the long4colheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the long4colheaderborder style there are no vertical rules.

Compare with the example-glossaries-user.tex test file, which has symbols for top level entries and most (but not all) child entries:

altlong4col-booktabs 🔗

The result is like the altlong4colheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the altlong4colheaderborder style there are no vertical rules.

longragged-booktabs 🔗

The result is like the longraggedheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the longraggedheaderborder style there are no vertical rules.

longragged3col-booktabs 🔗

If you are using glossaries-extra you may want to consider the long-name-desc-loc style instead.

The result is like the longragged3colheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the longragged3colheaderborder style there are no vertical rules.

altlongragged4col-booktabs 🔗

If you are using glossaries-extra you may want to consider the long-name-desc-sym-loc style instead.

The result is like the altlongragged4colheader style but with horizontal rules above and below the header and at the bottom of the table. Unlike the altlongragged4colheaderborder style there are no vertical rules.

glossary-super

The super styles use the supertabular environment, so the glossary-super package automatically loads the supertabular package.

Note that, in general, longtable is more reliable, but there are some situations (such as where TeX’s output routine is altered) where longtable doesn’t work.

Take care not to confuse these table styles with the mcols styles.

These styles are all analogous to the long styles.

super 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{lp{\glsdescwidth}}

\setglossarystyle{super}
  \renewenvironment{theglossary}%
    {\centering\tablehead{}\tabletail{}%
     \begin{supertabular}{lp{\glsdescwidth}}}%
    {\end{supertabular}}%

The result is essentially the same as for the long style.

This style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries. Note that this example has problematic page breaks. The long style performs much better.

In this case, the result isn’t exactly like the long style as there’s extra vertical spacing before the child entries and a spurious page break before the glossary. It’s better to use the long style instead, if possible.

superborder 🔗

The result is essentially the same as for the longborder style.

superheader 🔗

The result is essentially the same as for the longheader style.

superheaderborder 🔗

The result is essentially the same as for the longheaderborder style.

super3col 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{lp{\glsdescwidth}p{\glspagelistwidth}}

The result is essentially the same as for the long3col style.

This style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries. Note that this example has problematic page breaks. The long3col style performs much better.

In this case, the result isn’t exactly like the long3col style as there’s extra vertical spacing before the child entries and a spurious page break before the start of the glossary. It’s better to use the long3col style instead, if possible.

super3colborder 🔗

The result is essentially the same as for the long3colborder style.

super3colheader 🔗

The result is essentially the same as for the long3colheader style.

super3colheaderborder 🔗

The result is essentially the same as for the long3colheaderborder style.

super4col 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{llll}

Note that the test file loads example-glossaries-brief.tex, which doesn’t provide symbols for any of the dummy entries, so the symbol column is blank in the sample below. (The example-glossaries-symbols.tex test file has descriptions that are too long for this style.)

The result is essentially the same as for the long4col style.

This style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries.

The result is essentially the same as for the long4col style.

super4colborder 🔗

The result is essentially the same as for the long4colborder style.

super4colheader 🔗

The result is essentially the same as for the long4colheader style.

super4colheaderborder 🔗

The result is essentially the same as for the long4colheaderborder style.

altsuper4col 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{lp{\glsdescwidth}lp{\glspagelistwidth}}

To illustrate the way this style displays symbols, the test file can be modified to use example-glossaries-symbols.tex instead of example-glossaries-brief.tex. A vertical gap is inserted between letter groups, so use the nogroupskip option if your entries haven’t been sorted alphabetically.

The result is essentially the same as for the altlong4col style.

This style omits the name for sub-entries and is unsuitable for deeper than one-level sub-entries. The sample below uses the example-glossaries-childnoname.tex file with the subentrycounter option set, which automatically numbers the sub-entries. Note that this example has problematic page breaks. The altlong4col style performs much better.

In this case, the result isn’t exactly like the altlong4col style as there’s extra vertical spacing before the child entries and a spurious page break before the glossary. It’s better to use the long style instead, if possible.

altsuper4colborder 🔗

The result is essentially the same as for the altlong4colborder style.

altsuper4colheader 🔗

The result is essentially the same as for the altlong4colheader style.

altsuper4colheaderborder 🔗

The result is essentially the same as for the altlong4colheaderborder style.

glossary-superragged

\usepackage{glossary-superragged}

superragged 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{l>{\raggedright}p{\glsdescwidth}}

The result is essentially the same as for the longragged style.

superraggedborder 🔗

The result is essentially the same as for the longraggedheader style.

superraggedheader 🔗

The result is essentially the same as for the longraggedheader style.

superraggedheaderborder 🔗

The result is essentially the same as for the longraggedheaderborder style.

superragged3col 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{l>{\raggedright}p{\glsdescwidth}%
>{\raggedright}p{\glspagelistwidth}}

The result is essentially the same as for the longragged3col style.

superragged3colborder 🔗

The result is essentially the same as for the longragged3colborder style.

superragged3colheader 🔗

The result is essentially the same as for the longragged3colheader style.

superragged3colheaderborder 🔗

The result is essentially the same as for the longragged3colheaderborder style.

altsuperragged4col 🔗

\tablehead{}\tabletail{}%
\begin{supertabular}{l>{\raggedright}p{\glsdescwidth}l%
>{\raggedright}p{\glspagelistwidth}}

The result is essentially the same as for the altlongragged4col style.

altsuperragged4colborder 🔗

The result is essentially the same as for the altlongragged4colborder style.

altsuperragged4colheader 🔗

The result is essentially the same as for the altlongragged4colheader style.

altsuperragged4colheaderborder 🔗

The result is essentially the same as for the altlongragged4colheaderborder style.

glossary-bookindex

\usepackage{glossary-bookindex}

\usepackage[stylemods=bookindex]{glossaries-extra}

bookindex 🔗

There are a number of hooks that can be used to adjust the style. You can also use the post-name hook, but you can’t use the post-description hook as the description isn’t shown. The default number of columns is 2, but this may be changed by redefining \glsxtrbookindexcols. Each letter group is headed by default and, if hyperref has been loaded, the letter group will also be added to the PDF bookmarks.

The following glossary was created with the test file used to generate the documents in the above examples, which uses makeindex rather than bib2gls:

Each group header is centred (within the column) and shown in bold. The header is followed by \indexspace (indirectly via \glsxtrbookindexpostgroupskip). The group skip comes before the header (except for the first header) and can be suppressed with the nogroupskip option.

The test entries are all ASCII, so in this case there are no problems caused by extended characters in the letter group headings. The PDF bookmarks (as displayed in Okular) are shown below:

The exact style of the bookmarks varies according to the PDF viewer, but most have a tree-like format with collapsible elements.

Each top-level name is flush left, followed by a space and then the location list.

The location lists in the above document all simply contain “1” because each entry was indexed on page one by the command \glsaddall, which iterates over all the defined entries and performs \glsadd for each one. This just indexes without producing any text. Although this is useful in the test document to show where the location appears, the locations aren’t usually desired when using \glsaddall as it’s not helpful to the reader to have the same page appearing in every location list.

These types of iterative commands aren’t available with bib2gls as they require all the entries to be defined first. However, with bib2gls, the required entries aren’t defined on the first LaTeX run so the iterative commands have nothing to loop over. Instead, bib2gls can be instructed to select all entries defined in the designated .bib file with the selection option:

\GlsXtrLoadResources[
   selection=all, % select all entries in the given .bib file(s)
   src={entries1,entries2}% data given in entries1.bib and entries2.bib
]

In real documents, this is better than \glsaddall as it doesn’t automatically add an unnecessary location. This means that the closest matching test file (test-styles-bib2gls.tex) to the above doesn’t show any locations:

This now just shows the letter group headings and each entry’s name.

If I add the following line to the test file’s document environment:

\Gls{lorem} and \gls{ac}.

Now the “ac” and “lorem” terms are followed by a comma, a space and then the location list..

Another way of demonstrating the position of the locations is to fake a location for each entry with:

\forglsentries{\thisentry}{%
 \GlsXtrSetField{\thisentry}{location}{\glsnoidxdisplayloc{}{page}{glsnumberformat}{1}}%
}

With the standard \makeglossaries method, there is simply a space between the name and the location, but with bib2gls a comma is inserted if the location list is non-empty. If you are using bib2gls and don’t want any locations visible, then use the resource option save-locations=false rather than the package option nonumberlist. Otherwise you will end up with spurious commas.

The name is displayed using \glsxtrbookindexname{label}, which is simply defined as:

\newcommand*{\glsxtrbookindexname}[1]{\glossentryname{#1}}

\renewcommand*{\glsxtrbookindexname}[1]{%
 \glossentryname{#1} (\glsentrydesc{#1})%
}

To test if the entry has a description and to allow for the post-description hook and description-related category attributes, you can instead use:

\renewcommand*{\glsxtrbookindexname}[1]{%
 \glossentryname{#1}%
 \ifglshasdesc{#1}{ (\glossentrydesc{#1}\glspostdescription)}{}%
}

You can add other fields in a similar manner. Alternatively, you can use the post-name hook to make minor adjustments according to the entry’s category.

A more comprehensive view of the bookindex style can be obtained with the testidx-glossaries package (provided with testidx). The test file now implicitly loads glossaries-extra with the record option, so

\usepackage[bib2gls]{testidx-glossaries}

\usepackage[record]{glossaries-extra}

\usepackage{glossary-bookindex}
\setglossarystyle{bookindex}

\tstidxmakegloss[sort=sv]

The dummy text (which includes commands like \gls) is generated with \testidx and the glossary is displayed with: \tstidxprintglossary[title=Index] which, in this case, does \printunsrtglossary[title=Index]

Since this page is about the way the glossary is displayed (the style) rather than how the glossary is sorted and collated, the testidx commands aren’t particularly important.

This example uses the fancyhdr package to display the first and last word in the page footer. The page style is switched on with:

\fancypagestyle{plain}{%
  \fancyhead{}%
  \lfoot{\glsxtrbookindexfirstmark}%
  \cfoot{\thepage}%
  \rfoot{\glsxtrbookindexlastmark}%
}%
\pagestyle{fancy}%
\lhead{\thepage}%
\lfoot{\glsxtrbookindexfirstmark}%
\cfoot{}%
\rfoot{\glsxtrbookindexlastmark}%

(See the fancyhdr documentation for more information about the page style commands.) The two commands provided by the glossary-bookindex package used in the page style are \glsxtrbookindexfirstmark (which displays the first marked word for that page) and \glsxtrbookindexlastmark (which displays the last marked word for that page).

The top-level entries need to have the name added to the list of page marks (for access with the two commands described above):

\renewcommand{\glsxtrbookindexname}[1]{%
  \glsxtrbookindexmarkentry{#1}%
  \glossentryname{#1}%
}

\renewcommand{\glsxtrbookindexsubname}[1]{%
  \glossentryname{#1}%
}

These adjustments to the style require an extra LaTeX run, so the complete document build is:

pdflatex test-bookindex
bib2gls --group test-bookindex
pdflatex test-bookindex
pdflatex test-bookindex

The locations are in different colours to test the location encapsulator mechanism (the format key used in the optional argument of commands like \gls) and are not relevant to the style.

If you want to include descriptions in the style, you need to instruct the testidx-glossaries package to provide descriptions for the dummy entries with the desc option:

\usepackage[desc,bib2gls]{testidx-glossaries}

The name hook can be modified to include the description, as in the earlier example:

\renewcommand{\glsxtrbookindexname}[1]{%
  \glsxtrbookindexmarkentry{#1}%
  \glossentryname{#1}%
  \ifglshasdesc{#1}{: \glsentrydesc{#1}}{}%
}

If the sub-entries also need descriptions, then add similar code to \glsxtrbookindexsubname.

The indentation is governed by \glstreeitem (defined in glossary-tree.sty) and is also used by the index style. You can alter the indentation by redefining this command. For example:

\renewcommand{\glstreeitem}{\par\hangindent 20pt\relax}

The following example fakes the page numbers just for entries with a parent:

\forglsentries{\thisentry}{%
 \ifglshasparent{\thisentry}
 {\GlsXtrSetField{\thisentry}{location}{\glsnoidxdisplayloc{}{page}{glsnumberformat}{1}}}%
 {}%
}

It uses the example-glossaries-parent.bib example file.

This results in a comma and location list for each sub-entry, but not for the top-level entry.

This style supports sub-groups, which can be enabled with the group-level resource option:

\GlsXtrLoadResources[
 src={example-glossaries-parent},
 group-level=all
]

\BibGlsOptions{group}

The top-level group headings are centred, but the sub-headings have the same indent as their hierarchical level.

glossary-longextra

\usepackage{glossary-longextra}

\usepackage[stylemods=longextra]{glossaries-extra}

The styles in this package use longtable by default. If you have a single-paged glossary and you’d rather use tabular then you can switch with \GlsLongExtraUseTabulartrue before you set the style. If you switch, you can change the default vertical alignment for the tabular environment by redefining \glslongextraTabularVAlign.

There are four types of columns that may be displayed by these styles: name, description, symbol and location. The style name indicates which columns are shown by that style and their order. For example, long-name-desc has just two columns: name and description (no symbol or location list).

Version 1.49 introduced some extra styles that had columns for the short and long form of abbreviations, and version 1.50 introduced custom columns.

The column alignments are given by the commands: \glslongextraNameAlign (defaults to l), \glslongextraDescAlign (defaults to >{\raggedright}p{\glsdescwidth}), \glslongextraSymbolAlign (defaults to c), and \glslongextraLocationAlign (defaults to >{\raggedright}p{\glspagelistwidth}). These may be redefined as required.

Unlike the base long styles, these styles have hooks to calculate the value of \glsdescwidth (the width of the description column). To take advantage of these hooks, it’s necessary to provide the widest entry name. This can simply be done with the set-widest resource option when using bib2gls. Alternatively, you can use \glslongextraSetWidest{text}, where text is the widest name. If you don’t do this, the table may end up too wide or too narrow.

The column headers are encapsulated with \glslongextraHeaderFmt, which just uses \textbf by default to give bold headers. As with the base long styles, the header text is given by: \entryname (for the name column), \descriptionname (for the description column), \symbolname (for the symbol column) and \pagelistname (for the location list column).

Unlike the other long styles provided with just the base glossaries package, these styles have a hook to format the group headings, but it does nothing by default. It’s defined as:

\newcommand{\glslongextraGroupHeading}[2]{}

\renewcommand*{\glslongextraGroupHeading}[2]{%
  \glsxtrgetgrouptitle{#2}{\thisgrptitle}%
  \glslongextraHeaderFmt\thisgrptitle
  \tabularnewline*\noalign{\vskip\normalbaselineskip}%
}

There are commands that govern the way the name, description, symbol and location are displayed. The name is formatted with \glslongextraNameFmt{label}. There’s a bug in version 1.37 that’s been corrected in v1.38. The correct definition is:

\newcommand{\glslongextraNameFmt}[1]{%
 \glsentryitem{#1}\glstarget{#1}{\glossentryname{#1}}%
}

The description is formatted with \glslongextraDescFmt{label} which is defined as:

\newcommand{\glslongextraDescFmt}[1]{%
  \glossentrydesc{#1}\glspostdescription
}

The symbol is formatted with \glslongextraSymbolFmt{label} which is defined as:

\newcommand{\glslongextraSymbolFmt}[1]{\glossentrysymbol{#1}}

The location is formatted with \glslongextraLocationFmt{label}{location list} which is defined as:

\newcommand{\glslongextraLocationFmt}[2]{#2}

As with the base long styles, child entries are assumed to be homographs. The helper commands can be redefined to change this, but the tabular style will mask the hierarchical structure.

The name column for child entries is formatted with \glslongextraSubNameFmt{level}{label} but this just does the sub-entry counter (if enabled) and the hypertarget (if supported):

\newcommand{\glslongextraSubNameFmt}[2]{%
 \glssubentryitem{#2}\glstarget{#2}{\strut}%
}

The description column for child entries is formatted with \glslongextraSubDescFmt{level}{label} which defaults to the same as the top-level:

\newcommand{\glslongextraSubDescFmt}[2]{%
  \glslongextraDescFmt{#2}%
}

\newcommand{\glslongextraSubSymbolFmt}[2]{%
  \glslongextraSymbolFmt{#2}%
}

The location for child entries is formatted with \glslongextraSubLocationFmt{level}{label}{location list} which is defined as:

\newcommand{\glslongextraSubLocationFmt}[3]{#3}

long-name-desc 🔗

The result is very similar to the longragged-booktabs style but without the location list.

If the following code is added, then the group headings appear. If you are using bib2gls remember that you will need the --group (or -g) switch.

\renewcommand*{\glslongextraGroupHeading}[2]{%
  \glsxtrgetgrouptitle{#2}{\thisgrptitle}%
  \glslongextraHeaderFmt\thisgrptitle
  \tabularnewline*\noalign{\vskip\normalbaselineskip}%
}

The group header is just in the first column, in bold, followed by a vertical space of the same height as a normal line.

The following example has the subentrycounter package option and uses the example-glossaries-childnoname.bib sample file. Note that all child entries are treated the same way, regardless of their hierarchical level.

This puts the sub-entry counter in the name column (without the name).

I haven’t used the set-widest resource option in any of the above, but the default setting produces a fairly reasonable width. In the next example below, I’ve included the geometry package with the showframe option to show the table in relation to the available text area:

\usepackage[a4paper,showframe]{geometry}

The horizontal rules that span the entire table poke out into the right margin.

If I add the set-widest option to \GlsXtrLoadResources then the result shows a better fit:

long-name-desc-loc 🔗

The description width is calculated at the start of the style by \glslongextraLocSetDescWidth. This first uses \glslongextraSetDescWidth to calculate the width assuming just a name and description column, and then subtracts 2\tabcolsep+\glspagelistwidth, which is the width of the page list column and the width of the separator at either side of the column.

The result is very similar to the longragged3col-booktabs style.

long-desc-name 🔗

This is essentially the same as the long-name-desc style with the columns switched round.

long-loc-desc-name 🔗

This is essentially the same as the long-name-desc-loc style with the columns switched round.

long-name-desc-sym 🔗

The result is similar to the altlong4col-booktabs style, but without the page list column and the symbol is centred.

long-name-desc-sym-loc 🔗

The result is similar to the altlong4col-booktabs style (except that the symbol is centred).

long-name-sym-desc 🔗

This is essentially the same as the long-name-desc-sym style with the description and symbol columns switched round.

long-name-sym-desc-loc 🔗

This is essentially the same as the long-name-desc-sym-loc style with the description and symbol columns switched round.

long-sym-desc-name 🔗

This is essentially the same as the long-name-desc-sym style with the columns switched round.

long-loc-sym-desc-name 🔗

This is essentially the same as the long-name-desc-sym-loc style with the columns switched round.

long-desc-sym-name 🔗

This is essentially the same as the long-name-desc-sym style with the columns switched round.

long-loc-desc-sym-name 🔗

This is essentially the same as the long-name-desc-sym-loc style with the columns switched round.

long-sym-desc 🔗

This is like the long-name-desc style but the first column is headed “Symbol” and the symbol is shown instead of the name. Note that it’s left-aligned not centred.

The following adjustment also includes entries from the example-glossaries-brief.bib file:

\GlsXtrLoadResources[
 selection=all,
 src={example-glossaries-brief,
 example-glossaries-symbols}
]

In this example, the widest name is “consectetuer”, which is wider than the header “Symbol”. The table width can be automatically adjusted by adding the set-widest resource option and add the following redefinition:

\renewcommand{\glslongextraSymNoNameSetDescWidth}{%
 \glslongextraSetDescWidth
}

long-desc-sym 🔗

This is essentially the same as the long-sym-desc style with the columns switched round.

abbr-short-long 🔗

In this example, the example-glossaries-acronym-desc.bib is used, which has the description field set for all entries. I’ve used the long-short-desc abbreviation style:

\setabbreviationstyle[acronym]{long-short-desc}

As with the long-sym-desc style, the assumption is that the width of the column header will be wider that the width of the short form for each entry. The width of the second column is given by \glsdescwidth. Its value is calculated by \glslongextraShortNoNameSetDescWidth. In this example, I’ve used the long-short-desc abbreviation style, which means that the name will contain the long form as well as the short form, so using the widest name will be inappropriate.

abbr-long-short 🔗

In this example, the example-glossaries-acronym-desc.bib is used, which has the description field set for all entries. I’ve used the long-short-desc abbreviation style:

\setabbreviationstyle[acronym]{long-short-desc}

The default nogroupskip=false setting is on, which means there’s a vertical gap between letter groups.

This is essentially the same as the abbr-short-long style with the columns switched round.

long-name-custom1 🔗

\newcommand{\glslongextraCustomIHeader}{%
 \MFUsentencecase{\glslongextraCustomIField}}

The alignment for the name column is given by \glslongextraNameAlign, which is as for styles such as long-name-desc. The alignment for the user column is given by \glslongextraCustomIAlign. This defaults to l (left-aligned). If your custom field values are long and likely to need a paragraph cell, you will need to redefine this as appropriate.

This example uses the example-glossaries-user.bib test file, which is provided with glossaries-extra v1.50+, and corresponds to the example-glossaries-user.tex test file. This test set has all six user fields set for the top-level entries and a random selection set for child entries.

This results in a very narrow table as neither the names nor the user1 values are wide. Child entries don’t have the name display, which means that if the user1 field also isn’t set, it will result in blank rows.

long-custom1-name 🔗

long-name-custom2 🔗

The first two columns are as for the long-name-custom1 style. As above, this example uses the example-glossaries-user.bib test file.

long-custom2-name 🔗

long-name-custom3 🔗

The first three columns are as for the long-name-custom2 style. As above, this example uses the example-glossaries-user.bib test file.

long-custom3-name 🔗

long-name-custom1-desc 🔗

The description column is as for the long-name-desc style but the description width \glsdescwidth is calculated using \glslongextraCustomISetDescWidth. This first calculates the width according to \glslongextraSetDescWidth and then subtracts the width of the first custom column header. If the description width is incorrect, you will need to redefine \glslongextraCustomISetDescWidth to set \glsdescwidth to a more appropriate value. Alternatively, redefine \glslongextraDescAlign.

As above, this example uses the example-glossaries-user.bib test file.

long-name-custom2-desc 🔗

The description column is as for the long-name-desc style but the description width \glsdescwidth is calculated using \glslongextraCustomIISetDescWidth. This first calculates the width according to \glslongextraCustomISetDescWidth and then subtracts the width of the second custom column header. If the description width is incorrect, you will need to redefine \glslongextraCustomIISetDescWidth to set \glsdescwidth to a more appropriate value. Alternatively, redefine \glslongextraDescAlign.

As above, this example uses the example-glossaries-user.bib test file.

long-name-custom3-desc 🔗

The description column is as for the long-name-desc style but the description width \glsdescwidth is calculated using \glslongextraCustomIIISetDescWidth. This first calculates the width according to \glslongextraCustomIISetDescWidth and then subtracts the width of the third custom column header. If the description width is incorrect, you will need to redefine \glslongextraCustomIIISetDescWidth to set \glsdescwidth to a more appropriate value. Alternatively, redefine \glslongextraDescAlign.

As above, this example uses the example-glossaries-user.bib test file.

long-desc-custom1-name 🔗

long-desc-custom2-name 🔗

long-desc-custom3-name 🔗

glossary-topic

\usepackage{glossary-topic}

\usepackage[stylemods=topic]{glossaries-extra}

If you want to use this package with the set-widest resource option in bib2gls you will also need the glossary-tree package:

\usepackage[stylemods={tree,topic}]{glossaries-extra}

This package provides styles designed for glossaries that are lists of topics. That is, the top-level entries are considered topic titles (which may or may not have an associated symbol or description) and the sub-entries are items within that topic. By default the number list isn’t shown for the top-level entries but is shown after the description for sub-entries (unless suppressed with nonumberlist or save-locations=false).

The glossary-topic package automatically loads the multicol package.

topic 🔗

There’s a hook \glstopicMarker before the name, which does nothing by default, but can be redefined to add the topic to the PDF bookmarks or the table of contents. For example:

\renewcommand*{\glstopicMarker}[1]{%
 \pdfbookmark[1]{topic.#1}{\glsentryname{#1}}}

If the description is set, this is shown with the first letter converted to upper case and is followed by the post-description hook.

The sub-entries are indented from the second level onwards in a similar manner to the tree style. If you want an effect like the alttree style then you need to load the glossary-tree package to provide \glssetwidest.

There are no letter group separators with this style by default so the nogroupskip option has no effect. If you want letter groups as well, you can redefine \glstopicGroupHeading. For example:

\renewcommand*{\glstopicGroupHeading}[1]{%
  \glsxtrgetgrouptitle{#1}{\thisgrptitle}%
  \section*{\thisgrptitle}%
}

The example below was created with:

\usepackage
 [
   record,% using bib2gls
   stylemods={topic},% load glossary-topic.sty
   postdot% add post-description dot
 ]
 {glossaries-extra}

Adding glossary-tree.sty and the set-widest option produces a different result:

topicmcols 🔗

\renewcommand*{\glstopicColsEnv}{multicols*}

As with the topic style, the formatting changes if the set-widest option is used. Example without set-widest:

Example with set-widest:

glossary-table

\usepackage{glossary-table}

\usepackage[stylemods=table]{glossaries-extra}

This package doesn’t provide a general glossary style, but a single highly customised style that’s used by an accompanying command (also defined in glossary-table.sty) \printunsrttable that uses the style. The style should not be used in any other context and will be implicitly set by \printunsrttable (so don’t use the style option). Most of the options that can be passed to \printunsrtglossary can also be passed to \printunsrttable.

The nonumberlist setting is ignored, as the style doesn’t show the location list. (The location list is stored in the location field, unless you use the save-locations=false option, so it can be accessed if required.) Groups (heading and group skip) are off by default. If you want them, you need the groups option in \printunsrttable as well as invoking bib2gls with the --group switch. You will also need to load glossary-longbooktabs to support the group skip. Group headings will span across all columns of the table, which may leave some empty blocks at the end of the previous row.

Since \printunsrttable internally uses \printunsrtglossary, all entries will be listed in order of definition. Although this command is designed for use with bib2gls (which ensures that the order of definition matches the requested order), you can use \printunsrttable in addition to using \printglossary if the order of definition is suitable, but child lists and groups won’t be supported.

The purpose of \printunsrttable is to display the glossary as a table with the glossary title in the caption (not as a section title). What’s normally the table of contents title will be used in the list of tables. Any preamble or postamble provided will be incorporated into the table header or footer (spanning the table).

The aim is to allow multiple entries to appear on a single row. Consider the earlier long-name-custom1 style example, which had a very narrow table because the name and custom values were all quite short. A more compact table with multiple entries per row may be more desirable (particularly if print cost is an issue).

The nature of tabular-like environments, such as the longtable environment, makes it hard to have conditionals and increment a count register to keep track of when to start a new column and when to start a new row. Attempts to do this in a glossary style frequently cause errors, such as missing \omit or misplaced \noalign.

The \printunsrtglossary command has to construct the contents of the glossary by iterating over all entries. In order to avoid the tabular problems, it appends each entry item to an internal command until that command contains the entire content of the glossary. That means that the problematic loop and conditionals are performed outside of the tabular context. The \printunsrttable command makes use of the \printunsrtglossary hooks to insert the appropriate alignment or new row commands while the glossary content is being constructed. With par=justified or par=ragged the hooks can also be used to calculate the column widths.

Child entries are automatically filtered out, but level 1 entries can still be shown if you use the save-child-count resource option. In that case, they will be appended to the parent entry information. The child name isn’t shown by default, but if you want it displayed, use the following redefinition:

\renewcommand{\glstableSubNameFmt}[1]{#1}

If you want to apply any additional filtering, you can redefine \glstableiffilter. This takes three arguments. The first is the entry label. This command should do the second argument if the entry should be filtered and the third argument otherwise. Only top-level entries will be passed to this command. If child lists are visible, you can filter child entries by redefining \glstableiffilterchild, which has the same syntax as \glstableiffilter.

Each entry is displayed in a block. The block may be a single table cell or may comprise multiple columns. The block style determines how the entry is displayed and how many columns a single entry requires and where, if applicable, the child entries should be shown. The block style is set with the block-style option, and the number of blocks per row is set with the blocks option. The total number of columns in the table will be the number of columns per block multiplied by the number of blocks per row.

The following examples are all based on the file test-table.tex, which contains:

\documentclass{article}

\usepackage[T1]{fontenc}
\usepackage[stylemods=table,record]{glossaries-extra}

\GlsXtrLoadResources[
  src={example-glossaries-user},
 % resource options
  selection=all
]

\renewcommand{\glstableotherfield}{useriii}

\begin{document}
\printunsrttable[
 % table options
]
\end{document}

The default setting is blocks=2,block-style=name-desc which indicates two blocks per row using the name-desc block style. Predefined block styles are illustrated below using the default blocks=2 setting.

Block styles may show the name. In example-glossaries-user.bib, the name for top-level entries is in the form “letter name”, where letter is a lowercase letter. Child entry names are in the form “letter/n name”, where n is a number.

Block styles may show the description. In example-glossaries-user.bib, the description for top-level entries is in the form “letter description”, where letter is a lowercase letter. Child entry descriptions are in the form “child n of letter”, where n is a number.

Block styles may show the symbol. In example-glossaries-user.bib, the symbol, if set, will be a mathematical symbol, such as \alpha, possibly with a subscript or superscript. The symbols aren’t unique (for example, there are two top-level entries that both have symbol={\ensuremath{\lambda}}).

Block styles may support a custom “other” field. If required, the internal field label for this field should be in the definition of \glstableotherfield. The default definition is empty, which means no other field is required. The example document redefines this command to useriii, which corresponds to the user3 key. In example-glossaries-user.bib, this value, if set, is a lowercase Roman numeral. (Again, these aren’t unique.)

Block styles default to using horizontal rules provided by the booktabs package. If you don’t want them, use the rules=false option. There is a header row by default. If you don’t want it, use the header=false option. Each block is separated by a vertical line by default. If you don’t want it, use blocksep={}. If you don’t want a caption, use caption=false.

Each column within a block may be designated for a single field (such as the name) or may include other fields, if they are set. The header is determined by the primary field for the column. The associated commands are:

\glstablenameheader

The header for the name column. Default: \entryname

\glstabledescheader

The header for the description column. Default: \descriptionname

\glstablesymbolheader

The header for the symbol column. Default: \symbolname

\glstableotherheader

The header for the other column. Default: \MFUsentencecase{\glstableotherfield}

By default, the column alignment specifiers will be single-line specifiers: l (left), c (centred) or r (right) according to the column designation. If you need line-wrapping, use par=justified or par=ragged.

name 🔗

This shows the name followed by the symbol, other and description (if set) in parentheses. Note that the symbol is followed by a space but the other value is followed by a comma. If the other value isn’t present, the comma won’t be present either.

If you want the child entries, you will need the save-child-count resource option:

\GlsXtrLoadResources[
  src={example-glossaries-user},
  save-child-count,
  selection=all
]

\printunsrttable[
  block-style=name,
  par=ragged
]

By default, the child name isn’t shown.

name-desc 🔗

With the save-child-count option, the child list will be placed in the same column as the description. Remember that you will need to use par=ragged, otherwise the table will be too wide. The child list is like the child list for the name block style, but doesn’t include the symbol.

name-symbol 🔗

With the save-child-count option, the child list will be placed at the end of the name column (with par=false). You may need to use par=ragged (as in the example shown below), if the table ends up too wide. The child list shows the other value (if set) followed by the symbol, but doesn’t include the description.

Note that the child symbol is in the name column, not in the symbol column. This is because the child list is contained within a tabular environment in the same cell as the parent name.

desc-name 🔗

The result is like the name-desc block style example, but with the columns switched round.

symbol-name 🔗

The result is like the name-symbol block style example, but with the columns switched round.

The child list can be shown with the save-child-count resource option. Note that the child symbol comes before the other value.

name-symbol-desc 🔗

In this example, the table ends up too wide with the default par=false, so you may need to change that setting to allow line wrap with this block style.

The child list can be shown with the save-child-count resource option. The example below also uses par=ragged otherwise the table will be too wide. The child list is in the description column with the child symbol, other value (if set) and description. Note that the par=ragged setting also enables line wrap for the child entries.

name-other-desc 🔗

The child list can be shown with the save-child-count resource option. The example below also uses par=ragged otherwise the table will be too wide. The child list is in the description column with the child other value (if set) and description. Note that the par=ragged setting also enables line wrap for the child entries.

desc-other-name 🔗

The child list can be shown with the save-child-count resource option. The example below also uses par=ragged otherwise the table will be too wide. Note that the par=ragged setting also enables line wrap for the child entries.

The child list is in the description column with the child description and (if set) the other value. This is the reverse of the name-other-desc block style, which has the other value before the description.

name-symbol-other-desc 🔗

In this example, the table ends up too wide with the default par=false. However, if you change this to allow line wrapping, the columns will end up too narrow. The header text is also likely to be too long, so you may either need to suppress it with header=false or redefine the commands used to produce the header text (\glstablenameheader, \glstablesymbolheader, \glstableotherheader, and \glstabledescheader).

The child list can be shown with the save-child-count resource option, however this is likely to make the table too wide unless you have a wide page or unless you have just one block per row (blocks=1). The child list will be in the description column and is formatted like the child list for the name-symbol-desc block style.

name-desc-symbol 🔗

In this example, the table ends up too wide with the default par=false, so you may need to change that setting to allow line wrap with this block style.

The result is like the name-symbol-desc example, but with the symbol and description columns swapped.

With the save-child-count resource option, the child list will be in the description column with the child other value (if set), description and symbol. This is like the name-symbol-desc block style, but has the symbol after the description.

desc-symbol-other-name 🔗

In this example, the table ends up too wide with the default par=false, so you may need to change that setting to allow line wrap with this block style.

The result is like the name-symbol-other-desc example, but with the name and description columns swapped.

With the save-child-count resource option, the child list will be in the description column with the child description, symbol and (if set) the other value.

desc-other-symbol-name 🔗

In this example, the table ends up too wide with the default par=false, so you may need to change that setting to allow line wrap with this block style.

The result is like the desc-symbol-other-name example, but with the symbol and other columns swapped.

With the save-child-count resource option, the child list will be in the description column with the child description, the other value (if set) and the symbol.

name-other-symbol-desc 🔗

In this example, the table ends up too wide with the default par=false, so you may need to change that setting to allow line wrap with this block style.

The result is like the name-symbol-other-desc example, but with the symbol and other columns swapped.

With the save-child-count resource option, the child list will be in the description column with the child other value (if set), symbol and description.

name-other 🔗

With the save-child-count resource option, the child list will be in the other column and will just show the child other value. This will lead to blank rows if the other field isn’t set. (It’s possible to filter them out by redefining \glstableiffilterchild.)

other-name 🔗

other-symbol 🔗

With the save-child-count resource option, the child list will be in the other column and will show the child other (if set) and the symbol. Note that if both the other value and the symbol aren’t set, this will lead to blank rows. (It’s possible to filter them out by redefining \glstableiffilterchild.)

symbol-other 🔗

The result is like the other-symbol example, but the symbol and other columns are switched.