pg.migration.core — com.github.igrishaev/pg2-migration 0.1.17

Common migration functions.

-migrate clj

(-migrate scope id-migration's)

A general migrate forward function. Don't use it directly. Accepts the scope map and a seq of pairs (id, migration). Runs the migrations one by one. Tracks the applied migrations in the database.

-rollback clj

(-rollback scope id-migration's)

A general rollback function. Don't use it directly. Accepts the scope map and a seq of pairs (id, migration). Rolls back the migrations one by one. Removes the records from the database.

cleanup-slug clj

(cleanup-slug slug)

Turn a slug fragment into a human-friendly text.

create-migration-files clj

(create-migration-files migrations-path)

(create-migration-files migrations-path {:keys [id slug]})

Create a couple of new migration files, prev and next. When not set, the id is generated from the current UTC time. Return a pair of File objects, prev and next.

DATETIME_PATTERN clj

DEFAULTS clj

ensure-table clj

(ensure-table conn migrations-table)

Having a db connection and the name of the migrations table, create a table if it doesn't exist.

error! clj macro

(error! template & args)

generate-datetime-id clj

(generate-datetime-id)

Generate a Long number based on the current UTC datetime of a format YYYYmmddHHMMSS, e.g. 20240308154432.

get-applied-migration-ids clj

(get-applied-migration-ids conn table)

Read all the applied migrations from the database and return a set of their ids.

group-parsed-urls clj

(group-parsed-urls parsed-urls)

Having a seq of maps (parsed URLs), unify them by id. Return a sorted map, an instance of clojure.lang.Sorted.

letters clj

(letters n)

(letters n ch)

Repeat a certain string/character N times.

log-connection-info clj

(log-connection-info config)

Report the connection information.

log-sql clj

(log-sql url sql)

make-file-name clj

(make-file-name id text direction)

Compose a file name based on the id, an optional slug, and the direction type.

make-scope clj

(make-scope config)

Having a raw config map, do the following:

• apply config defautls
• build migrations map
• get applied migration IDs from the database
• mark applied migrations;
• detect the current migration;

Return a map of all these fields.

migrate-all clj

(migrate-all config)

Migrate all the pending forward migrations.

migrate-one clj

(migrate-one config)

Apply the next single migration.

migrate-to clj

(migrate-to config id-to)

Migrate forward to a certain migration by its ID.

parse-name clj

(parse-name file-name)

parse-url clj

(parse-url url)

Split a URL pointing to a migration file to various fields.

RE_FILE clj

read-disk-migrations clj

(read-disk-migrations path)

Read all the migrations from a resource.

rollback-all clj

(rollback-all config)

Rollback all the previous migrations.

rollback-one clj

(rollback-one config)

Rollback the current migration.

rollback-to clj

(rollback-to config id-to)

Rollback to a certain migration by its ID.

text->slug clj

(text->slug text)

Turn a 'Human Friendly Text ' into something like 'human-friendly-text' to be a part of a migration file.

url->migrations clj

(url->migrations url)

Fetch all the migrations from the top-level URL. Return a sorted map like (id => migration-map).

validate-conflicts! clj

(validate-conflicts! migrations)

Try to find a situation when a migration with less ID was applied before another migration with a greater ID.

For example:

id applied 1 yes 2 yes 3 no 4 yes

Above, 3 is less then 4 but 4 has been applied whereas 3 has not. Usually it happens when two features get merged at the same time. To fix it, rename the migration 3 to 5 as follows:

id applied 1 yes 2 yes 4 yes 5 no

validate-duplicates! clj

(validate-duplicates! parsed-urls)

Having a seq of parsed URLs, find those that have the same (id, direction) pair. If found, an exception is thrown.

validate-migration-id! clj

(validate-migration-id! migrations id)

Check if the migration map has a certain id.