sqitch-add - Add a database change to plans
sqitch add widgets [options sqitch add blankets --all sqitch add --change sprockets pg sql sqitch add slinkies --require sprockets --set schema=industry
This command adds a database change to one or more plans. This will result in
the creation of script files in the deploy, revert, and verify directories,
and possibly others. The content of these files is determined by the
evaluation of templates. By default, system templates in
$(prefix)/etc/sqitch/templates are used. These can be overridden by a
single user by creating templates in
~/.sqitch/templates/ See “Templates”
The paths and extensions of the generated scripts depend on the configuration of Sqitch targets, engines, and the core. See sqitch-configuration for details.
Note that the name of the new change must adhere to the rules as defined in sqitchchanges.
By default, the
add command will add the change to the default plan and the
scripts to any top directories for that plan, as defined by the core
configuration and command-line options. This works well for projects in which
there is a single plan with separate top directories for each engine, for
example. Pass the
--all option to have it iterate over all known plans and
top directories (as specified for engines and targets) and add the change to
To specify which plans and top directories to which the change and its scripts
will be added, pass the target, engine, or plan file names as arguments. Use
--change to disambiguate the tag and change names from the other parameters
if necessary (or preferable). See “Examples” for examples.
The name of the change to add. The name can be specified with or without this option, but the option can be useful for disambiguating the change name from other arguments.
Name of a change that is required by the new change. May be specified multiple times. See sqitchchanges for the various ways in which changes can be specified.
Name of a change that conflicts with the new change. May be specified multiple times. See sqitchchanges for the various ways in which changes can be specified.
Add the change to all plans in the project. Cannot be mixed with target, engine, or plan file name arguments; doing so will result in an error. Useful for multi-plan projects in which changes should be kept in sync. Overrides the value of the
--no-allto override a true
A brief note describing the purpose of the change. The note will be attached to the change as a comment. Multiple invocations will be concatenated together as separate paragraphs.
For you Git folks out there,
Set a variable name and value for use in the templates. The format must be
--set comment='This one is for you, babe.'.
Location to look for the templates. If none is specified,
addwill first look in
~/.sqitch/templates/for each template, and fall back on
Name of the templates to use for the scripts. When Sqitch searches the template directory for templates, it uses this name to find them in subdirectories named for the various types of scripts, including:
Any templates found with the same name in additional subdirectories will also be evaluated.
This option allows one to define templates for specific tasks, such as creating a table, and then use them for changes that perform those tasks. Defaults to the name of the database engine (
Specify the path to a template for a specific type of script. Defaults to the individual templates and using
--template-name, found in
--template-directoryand the configuration template directories.
Specify a type of template to generate or not generate.
Open the generated change scripts in an editor.
Do not open the change scripts in an editor. Useful when
Path to the deployment plan file. Overrides target, engine, and core configuration values. Defaults to
Add a change to a project and be prompted for a note.
sqitch add widgets
Add a change and specify the note.
sqitch add sprockets --note 'Adds the sprockets table.'
Add a change that requires the
users change from earlier in the plan.
sqitch add contacts --requires users -n 'Adds the contacts table'
Add a change that requires multiple changes, including the change named
extract from a completely different Sqitch project named
sqitch add coffee -r users -r utilities:extract -n 'Mmmmm...coffee!'
Add a change that uses the
createtable templates to generate the scripts,
as well as variables to be used in that template (See
https://justatheory.com/2013/09/sqitch-templating/ for a custom template
sqitch add corp_widgets --template createtable \ -s schema=corp -s table=widgets \ -s column=id -s type=SERIAL \ -s column=name -s type=TEXT \ -s column=quantity -s type=INTEGER \ -n 'Add corp.widgets table.'
Add a change only to the plan used by the
vertica engine in a project:
sqitch add --change logs vertica -n 'Adds the logs table to Vertica.'
Add a change to just two plans in a project, and generate the scripts only for those plans:
sqitch add -a coolfunctions sqlite.plan pg.plan -n 'Adds functions.'
Sqitch contains a very simple set of templates for generating the deploy,
revert, and verify scripts, and you can create more of your own. By default,
Sqitch uses system-wide templates installed in
sqitch --etc-path to find out
where, exactly (e.g.,
$(sqitch --etc-path)/templates). Individual templates
may be overridden on a user basis by copying templates to
~/.sqitch/templates and making modifications. They may also be overridden
by using the
--template-name options, as well as
the template-specific options.
Sqitch looks for templates in the following directories, and in this order:
Each should consist of subdirectories named for the types of scripts to be
generated. These should include
verify, but you
can create any number of other directories to create additional scripts that
will end up in a directory of the same name.
Each directory should include one or more files ending in
main part of the file name can be anything, but by default Sqitch will
look for a file named for the database engine. Use the
to have Sqitch use a different file.
For example, say you have this directory structure:
templates/deploy/pg.tmpl templates/deploy/create_table.tmpl templates/revert/pg.tmpl templates/revert/create_table.tmpl templates/test/pg.tmpl templates/verify/pg.tmpl templates/verify/create_table.tmpl
Assuming that you’re using the PostgreSQL engine, the code for which is
when you add a new change like so:
sqitch add schema -n 'Creates schema'
Sqitch will use the
pg.tmpl files to create the following files in the
top directory configured for the project (See sqitch-configuration for
deploy/schema.sql revert/schema.sql test/schema.sql verify/schema.sql
If you want to use the
create_table templates, instead, use the
--template option, like so:
sqitch add user_table --template create_table -n 'Create user table'
Sqitch will use the
create_table.tmpl files to create the following files
in the top directory configured for the project (See sqitch-configuration
deploy/user_table.sql revert/user_table.sql verify/user_table.sql
Note that the
test file was not created, because no
test/create_table.tmpl template file exists.
The syntax of Sqitch templates is the very simple language provided by Template::Tiny, which is limited to:
This is the directive syntax. By default, the return value of the expression is output:
-- Deploy [% project %]:[% change %] to [% engine %]
You can add
-to the immediate start or end of a directive tag to control the whitespace chomping options:
[% IF foo -%] # remove trailing newline We have foo! [%- END %] # remove leading newline
[% IF %]
[% IF %] / [% ELSE %]
[% UNLESS %]
[% IF transactions %] BEGIN; [% ELSE %] -- No transaction, beware! [% END %]
[% FOREACH item IN list %]
Loop over a list of values:
[% FOREACH item IN requires -%] -- requires: [% item %] [% END -%]
If this is not sufficient for your needs, simply install Template::Toolkit and all templates will be processed by its more comprehensive features. See the complete Template Toolkit documentation for details, especially the syntax docs
Sqitch defines five variables for all templates. Any number of additional
variables can be added via the
--set option, like so:
sqitch add --set transactions=1 --set schema=foo
Any number of variables may be specified in this manner. You may then use
those variables in custom templates. Variables that appear multiple times will
be passed to the templates as lists of values for which you will likely want
[% FOREACH %]. If the templates do not reference your variables,
they will be ignored. Variables may also be specified in a
config section (see “Configuration Variables”). Variables
--set will override configuration variables.
The five core variables are:
The name of the change being added.
The name of the engine for which the change was added. One of
The name of the Sqitch project to which the change was added. The project name is set in the plan by the “
A list of required changes as passed via one or more instances of the
A list of conflicting changes as passed via one or more instances of the
Add the change to all the plans in the project. Useful for multi-plan projects in which changes should be kept in sync. May be overridden by
--no-all, or target, engine, and plan file name arguments.
Directory in which to find the templates. Any templates found in this directory take precedence over user- or system-specific templates, and may in turn be overridden by the
Name used for template files. Should not include the
.tmplsuffix. Overrides the default, which is the name of the database engine, and may in turn be overridden by the
Location of templates of different types. Core templates include:
But a custom template type can have its location specified here, as well, such as
add.template.unit_test. May be overridden by
A section defining template variables. Useful if you’ve customized templates with your own variables and want project-, user-, or system-specific defaults for them.
Boolean indicating if the add command should spawn an editor after generating change scripts. When true, equivalent to passing
--edit. Defaults off.
Part of the sqitch suite.