|
NAME | SYNOPSIS | DESCRIPTION | SYNTAX | EXAMPLES | SEE ALSO | COLOPHON |
|
|
|
RPM-QUERYFORMAT(7) Miscellaneous Information Manual RPM-QUERYFORMAT(7)
rpm-queryformat - RPM query mini-language
Query format
[PLACEHOLDER|ITERATOR|EXPRESSION|TEXT ...]
Placeholder
%[WIDTH]{TAG[:FORMAT]}
Iterator
[QUERYFORMAT ...]
Note: Square brackets are literal and QUERYFORMAT must contain an
array placeholder.
Expression
%|TAG?{PRESENT}:{ABSENT}|
Note: Vertical bars are literal, PRESENT and ABSENT are query
formats.
RPM supports a simple query language for extracting header data to
plain text. Its syntax resembles that of printf(3) where header
tags enclosed in %{ and }, for example %{NAME}, are used as
placeholders. A string written in this syntax is called a query
format.
Query formats are currently accepted in the following places:
1. Query operations with rpm(8), as a way to control the output
(see --queryformat)
2. rpm-scriptlets(7), to access arbitrary header data at runtime
(see -q)
3. Configuration macros in rpm-config(5) and rpmbuild-config(5),
such as %_query_all_fmt or %_rpmfilename
See the EXAMPLES section for more details.
A valid query format is composed of optional literal text and zero
or more placeholders. A placeholder has the following syntax:
%[WIDTH]{TAG[:FORMAT]}
When RPM produces output according to a query format, it
substitutes each placeholder with the TAG data from the queried
package and formats the data according to the specified WIDTH and
FORMAT (if given). The rest of the query format is copied to the
output unchanged.
Special characters such as \n or \t, with the exception of \0, are
also supported. Curly braces, the percent sign and the backslash
may be escaped with a backslash.
The individual components of a placeholder, as well as some
additional language features, are described in the subsections
below.
Tags
Every RPM package contains a key-value store (the header) that
represents the package's properties, such as its name, version or
file list. The keys in this store are called tags, and can be used
in a query format as placeholders for the values (data) they point
to. The data is either a single scalar (string, integer or binary
blob), or an array (argv) of multiple scalars of the same type.
To obtain a list of all tags supported by the installed version of
RPM, run the following command:
rpm -v --querytags
This will print a list of available tags, along with their
internal numbers and types, for example:
BASENAMES 1117 argv
BUILDHOST 1007 string
BUILDTIME 1006 int32
DESCRIPTION 1005 i18nstring
EPOCH 1003 int32
INSTALLTIME 1008 int32
NAME 1000 string
RELEASE 1002 string
SIZE 1009 int32
SUMMARY 1004 i18nstring
VERSION 1001 string
Tags are case-insensitive. Each tag also has an implicit variant
with the RPMTAG_ prefix, such as RPMTAG_NAME. Both variants can be
used in a query format interchangeably.
See the tags documentation in SEE ALSO for more details.
Formats
Each tag is printed in a specific output FORMAT. When no format is
specified, the default :string is used which displays a bare-bones
representation of the data. You can specify a different format by
appending its name after a colon, for example:
%{INSTALLTIME:date}
This is useful for tags that are not stored in human-readable
form, such as sizes, dates or signatures. Some formats allow for
performing more complex transformations of the data, such as
encoding it in base64 or wrapping it in JSON syntax.
The following formats are currently available:
Name Description
:armor Wrap a public key in ASCII armor
:arraysize Display number of elements in array tags
:base64 Encode binary data using base64
:date Use strftime(3) "%c" format
:day Use strftime(3) "%a %b %d %Y" format
:depflags Format dependency comparison operator
:deptype Format dependency type
:expand Perform macro expansion
:fflags Format file flags
:fstate Format file state
:fstatus Format file verify status
:hashalgo Display hash algorithm name
:hex Format in hexadecimal
:octal Format in octal
:humaniec Human readable number (in IEC 80000) where K =
1024, M = 1048576, ...
:humansi Human readable number (in SI) where K = 1000, M =
1000000, ...
:json Wrap data in JSON
:perms Format file permissions
:pgpsig Display signature fingerprint and time
:shescape Escape single quotes for use in a script
:string Display string format (default)
:tagname Display tag name
:tagnum Display tag number
:triggertype Display trigger suffix
:vflags Format file verification flags
:xml Wrap data in simple XML markup
Note that some formats are type specific, which means they can
only be used with a specific tag type. If you attempt to use them
with an incompatible type, RPM will replace them with a
placeholder text. For example, the following query format will
result in the output "(not a number)":
%{NAME:date}
Similarly, if a tag is not present in the queried package, it will
be replaced with the text "(none)".
Width
You can specify the minimum WIDTH of the data when printed, by
placing a number in between the % and { characters, similarly to
field width used in printf(3). Optionally, prefix the number with
the minus sign to left-justify the text.
Iterators
Tags pointing to arrays may be expanded to the individual elements
by using iterators. An iterator is a query (sub-)format enclosed
in square brackets that contains a placeholder for an array, for
example:
[%{FILENAMES}\n]
RPM will loop through the enclosed array, outputting the
(sub-)format each turn and substituting the placeholder with the
current element. This example includes a newline character, which
will result in one filename per line.
Multiple arrays may be placed into a single iterator. RPM will
then iterate over all of them in lockstep, meaning that the first
iteration will substitute each placeholder with its array's first
element, the second iteration will use each array's second
element, and so on. For example:
[%10{FILESIZES} %{FILENAMES}\n]
This is useful with parallel arrays that RPM keeps internally.
These are sets of arrays where all elements at the same index
correspond to each other. One can think of parallel arrays as
table columns that can be arbitrarily rearranged. The above
example will output a table of files and their sizes.
Sometimes, it is useful to combine an array with one or more
scalars in the same iterator. This is possible since RPM
internally treats all scalars as arrays of size one. However,
iterators may only contain arrays of the same size, otherwise RPM
will complain. To avoid that, you can "lock" an array so that only
its first element is used in each iteration, by prefixing the tag
with the equal sign, for example:
[%{=NAME} %{FILENAMES}\n]
This will produce an "annotated" file list where each filename is
preceded by the name of the package that contains it.
Expressions
Simple conditionals may be evaluated through query expressions in
the form of
%|EXPR|
where EXPR is an expression.
The only type of expression currently supported is a C-like
ternary operator for simple if/then/else conditions, which has the
syntax
%|TAG?{PRESENT}:{ABSENT}|
where PRESENT and ABSENT are query formats that should be expanded
if the queried package does and doesn't have the TAG,
respectively. Thus, expressions can be nested.
Example 1. Query commands
rpm -qa --queryformat "%-30{NAME} %{SIZE:humaniec}\n"
Print a table of all installed package names and their human-
readable sizes.
rpm -q --queryformat "%{NAME} %{INSTALLTIME:date}\n" fileutils
Print the fileutils package name followed by its installation
date.
rpm -q --queryformat "[%{FILEMODES:perms} %{FILENAMES}\n]" rpm
Print all filenames owned by the rpm package, one per line,
with each name preceded by a file permissions string similar
to the long listing format of ls(1).
rpm -q --queryformat "[%{REQUIRENAME} %{REQUIREFLAGS:depflags}
%{REQUIREVERSION}\n]" vlock
Print all capabilities on which the vlock package depends in
the form of version comparisons, one per line, with the
capability name and the required version as the left and right
operand, respectively.
rpm -qa --queryformat "%-30{NAME} %|PREINPROG?{ %{PREINPROG}}:{
no}|\n"
Print a table of all installed package names in the left
column and the program names used for their prein scriptlets
(such as the default /bin/sh) in the right column, with the
latter being no if the given package has no prein scriptlet.
Example 2. Command aliases
/usr/lib/rpm/rpmpopt-VERSION
A popt(3) alias file used by rpm(8) itself to implement
various shorthand options (such as --scripts) with the help of
query formats. Replace VERSION with the version of RPM
installed on your system.
Example 3. Transaction scriptlets
Print the exact file list of a package after its installation:
%post -q
for f in [%%{instfilenames} ]; do
echo $f
done
Note that the trailing space inside the square brackets is
required in order for the query format to be expanded to separate
words that the for statement can loop over.
See rpm-scriptlets(7) for more information on scriptlets in
general.
rpm(8), rpm-common(8), rpm-config(5), rpm-scriptlets(7),
rpm-spec(5), popt(3)
https://rpm.org/docs/latest/manual/tags
This page is part of the rpm (RPM Package Manager) project.
Information about the project can be found at
⟨https://github.com/rpm-software-management/rpm⟩. It is not known
how to report bugs for this man page; if you know, please send a
mail to man-pages@man7.org. This page was obtained from the
project's upstream Git repository
⟨https://github.com/rpm-software-management/rpm.git⟩ on
2026-01-16. (At that time, the date of the most recent commit
that was found in the repository was 2026-01-15.) If you discover
any rendering problems in this HTML version of the page, or you
believe there is a better or more up-to-date source for the page,
or you have corrections or improvements to the information in this
COLOPHON (which is not part of the original manual page), send a
mail to man-pages@man7.org
RPM 6.0.90 2026-01-16 RPM-QUERYFORMAT(7)
Pages that refer to this page: rpmspec(1), rpmbuild-config(5), rpm-config(5), rpm-scriptlets(7), rpm(8)