Incoming links: Sequelize transaction retry
@cirosantilli/_file/nodejs/sequelize/nodejs/sequelize/parallel_select_and_update.js Updated 2024-12-15 +Created 1970-01-01
This example is the same as nodejs/sequelize/raw/parallel_select_and_update.js, but going through Sequelize rather than with Sequelize raw queries.
NONE is not supported for now to not have a transaction at all because lazy.The examples illustrates: stackoverflow.com/questions/55452441/for-share-and-for-update-statements-in-sequelize
Sample invocation:where:
node --unhandled-rejections=strict ./parallel_select_and_update.js p 10 100 READ_COMMITTED UPDATEREAD_COMMITTED: one of the keys documented at: sequelize.org/master/class/lib/transaction.js~Transaction.html which correspond to the standard sQL isolation levels. It not given, don't set one, defaulting to the database's/sequelize's default level.UPDATE: one of the keys documented at: sequelize.org/master/class/lib/transaction.js~Transaction.html#static-get-LOCK. Update generates aSELECT FOR UPDATEin PostgreSQL for example. If not given, don't use anyFOR xxxexplicit locking.
Other examples:
node --unhandled-rejections=strict ./parallel_select_and_update.js p 10 100 READ_COMMITTED UPDATE
Then, the outcome is exactly as described at: nodejs/sequelize/raw/parallel_select_and_update.js:
READ_COMMITTED: failsREAD_COMMITTED UPDATE: worksREPEATABLE_READ: works, but is a bit slower, as it does rollbacksThis case also illustrates Sequelize transaction retries, since in this transaction isolation level transactions may fail:
- foreign keys are capitalized:
- you must give
foreignKeywhen using aliases, otherwise it fails subtely. That would be derived automatically. - stackoverflow.com/questions/41502699/return-flat-object-from-sequelize-with-association can't auto-flatten to reuse the database's
ORDER limitandoffsetdon't work withoutsubQuery: falsewhen doing includes! It is just too buggy. Examples of this can be found e.g. under nodejs/sequelize/many_to_many_same_model.js.- stackoverflow.com/questions/34059081/how-do-i-reference-an-association-when-creating-a-row-in-sequelize-without-assum hard to not duplicate foreign keys values everywhere
- stack traces permanently broken or requiring non-obvious configs:
- does not automatically update fields on hooks: github.com/sequelize/sequelize/issues/8586#issuecomment-422877555
- cannot change columns when other columns have constraints due to the backup table?
- you have to use
.get()forattributealiased fields, why? stackoverflow.com/questions/32649218/how-do-i-select-a-column-using-an-alias/69890944#69890944 .idgets added toSELECTno matter what, breakingGROUP BYunless you do horrible workarounds:- no simple built-in mechanism for transaction retries: Sequelize transaction retries
- impossible to do subqueries in general. Docs just tell you to use literals. This in particular prevents single query deletes with join as done at nodejs/sequelize/raw/many_to_many.js:Also, you can't get query strings either: github.com/sequelize/sequelize/issues/2325
- sequelize.org/master/manual/sub-queries.html: the docs actually just tell you to use literals, lol
- stackoverflow.com/questions/45354001/nodejs-sequelize-delete-with-nested-select-query
- migrations. Generally speaking, anything but the simplest migrations are exceedingly hard to get right, as you have to go very low level when doing migrations. Syntax can be very different from regular DB operations.
- no way to do (non-raw) queries during migrations, e.g. to update fields based on other fields in a complex way?
- github.com/sequelize/cli/issues/862
- stackoverflow.com/questions/18742962/add-data-in-sequelize-migration-script
- stackoverflow.com/questions/38671483/sequelize-migration-update-model-after-updating-column-attributes
- stackoverflow.com/questions/38998397/can-i-use-sequelize-models-in-migration-scripts
- stackoverflow.com/questions/45286429/custom-query-on-sequelize-seeder`queryInterface.sequelize.models` contains only
SequelizeMeta. Not sure why they have this limitation.
Edit: actually things will likely just work if immediately after making table changes you just instantiate a new sequelize and do any data changes. - stackoverflow.com/questions/56043246/node-js-sequelize-no-primary-keys-when-migrating/56046101#56046101
- SQLite
changeColumnmigrations do on delete cascades of other tables. SQLite does not have change column statements, so they have to drop and recreate tables, but they don't temporarily remove cascades, so you lose data: stackoverflow.com/questions/62667269/sequelize-js-how-do-we-change-column-type-in-migration/70486686#70486686 - associations require full explicit index construction: stackoverflow.com/questions/39651853/how-to-create-join-table-with-foreign-keys-with-sequelize-or-sequelize-cli
- ability to iterate over a large result without blowing up memory and without using limit + offset (which is inneficient e.g. when looping over recursive queries). This is also known as cursor or streaming interfaces:E.g. the Python SQLite interface supports this just fine: stackoverflow.com/questions/29582736/python3-is-there-a-way-to-iterate-row-by-row-over-a-very-large-sqlite-table-wi
- stack overflow
- stackoverflow.com/questions/28787889/how-can-i-set-up-sequelize-js-to-stream-data-instead-of-a-promise-callback
- stackoverflow.com/questions/43964067/how-to-implement-cursor-pagination-using-sequelize
- stackoverflow.com/questions/57164242/perform-sequelize-findall-in-a-huge-array
- stackoverflow.com/questions/55191891/how-to-loop-through-result-in-sequelize generic loop
- issue tracker
- stack overflow
- empty
attributes: []breaks some nested queries: github.com/sequelize/sequelize/issues/16436 - does not expose a iteration API that supports large arrays?E.g. Python SQLite does: stackoverflow.com/questions/29582736/python3-is-there-a-way-to-iterate-row-by-row-over-a-very-large-sqlite-table-wi