Add @effect/sql-pglite package by evelant · Pull Request #4692 · Effect-TS/effect

evelant · 2025-04-03T17:39:33Z

Type

Description

Add @effect/sql-pglite package. An @effect/sql driver for PGlite, single user postgresql in the browser or node.

Thanks for the feedback! I've had to put down my side project using pglite for now but as soon as I can find time I'll update the PR. As for the naming it's extra confusing because PGlite uses PGlite not Pglite or PgLite. I think my fingers just typed PgLite out of habit haha. I think you're right, probably best to change it to Pglite to match Sqlite. Ah I see how the formatting got a little off too -- prettier plugin in vscode auto-formatting on save.

evelant · 2025-06-03T22:08:31Z

I finally got around to cleaning this up. Sorry for the long delay!

nikelborm · 2025-06-04T10:41:51Z

+    const pgDumpMigrations = runPgDump(["--column-inserts", "--data-only", `--table=${table}`])
+
+    const pgDumpAll = Effect.map(
+      Effect.all([pgDumpSchema, pgDumpMigrations], { concurrency: 2 }),


Will setting concurrency: 2 here have any effect? Isn't dumping the database a blocking operation?

The PgliteMigrator is a direct copy of the PgMigrator just with commands switched out for Pglite instead of Pg. I'm not sure why the original PgMigrator does it this way TBH.

evelant · 2025-06-18T17:33:55Z

Any more changes needed or can this be merged now?

nikelborm · 2025-06-30T17:34:51Z

I think you probably have to ask Tim for a review manually. Because on my PR, where I had a lot of changes to SQL packages, he was automatically attached as per the CODEOWNERS file, while it seems this automation doesn't work when you are adding a completely new SQL-related package. Maybe you have to even add a new row there

nikelborm · 2025-07-08T10:04:11Z

Fixed failing CI in evelant/effect#3

evelant · 2025-07-22T22:30:21Z

@tim-smart Whenever you get a chance I'd appreciate your feedback on this so I can make any necessary changes, thanks!

tim-smart · 2025-07-23T04:49:12Z

+      executeValues(sql: string, params: ReadonlyArray<Primitive>) {
+        // PGlite doesn't have a values() method like postgres.js
+        // We'll just return the regular query results
+        return this.execute(sql, params, (r) => Object.values(r))


I think you will need an .map here? The results will be an array?

Maybe I was confused about what executeValues should return. Is it supposed to be an array of arrays of values? That's what this should do since the function param is the row transform, it transforms each row object into an array of values.

r will be an array of objects here, so you will need to map over each item.

tim-smart · 2025-07-23T04:49:44Z

+      ) {
+        // PGlite doesn't have a cursor method like postgres.js
+        // We'll fetch all results at once and convert to a stream
+        return Stream.fromEffect(


You can use Stream.fromIterableEffect

tim-smart · 2025-07-23T04:51:13Z

+      constructor(private readonly pg: PGlite) {}
+
+      private run(query: Promise<any>) {
+        return Effect.async<ReadonlyArray<any>, SqlError>((resume) => {


You can switch to tryPromise here

tim-smart · 2025-07-23T04:52:34Z

+        extensions: options.extensions ? (client as any) : ({} as any),
+        listen: (channel: string) =>
+          Stream.asyncPush<string, SqlError>((emit) =>
+            Effect.tryPromise({


You would use Effect.acquireRelease here. I think this would immediately unsubscribe.

tim-smart · 2025-07-23T04:52:50Z

+          Effect.tryPromise({
+            try: () => client.query(`NOTIFY ${channel}, '${payload}'`),
+            catch: (cause) => new SqlError({ cause, message: "Failed to notify" })
+          }).pipe(Effect.map(() => void 0))


Suggested change

}).pipe(Effect.map(() => void 0))

}).pipe(Effect.asVoid)

tim-smart · 2025-07-23T21:34:15Z

+
+test("should work", () => expect(true))
+describe("PgliteClient", () => {


Suggested change

test("should work", () => expect(true))

describe("PgliteClient", () => {

describe("PgliteClient", () => {

bwbuchanan · 2025-11-04T13:22:16Z

This is especially needed now that @effect/sql-pg uses pg instead of postgres, and has lost support for the socket configuration option, which I was previously using to integrate with pglite for integration-testing database code.

evelant · 2025-11-04T14:50:18Z

Sorry I dropped the ball on finishing this PR. I'll pick it up again ASAP.

thewilkybarkid · 2025-11-27T15:14:34Z

Anything we could help with? 🙂

Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Tim Smart <hello@timsmart.co>

evelant · 2025-12-17T14:06:31Z

Sorry for taking forever to get back to this. I've been extremely busy. I cleaned everything up, addressed the review comments, rebased on main, and updated everything to match latest changes in sql. @tim-smart if you would take another look I'd appreciate it!

evelant · 2025-12-17T15:10:19Z

Sign in to view

hmm, I'm not sure how that happened! Thanks for pointing it out, I'll try to clean it up, although I'm not sure of the best way to do so. Maybe I should copy the pglite changes to a new branch and open a new PR that's clean.

Hmm yeah IDK what happened here. My history looks clean locally. The commits relating to pglite are clean. It seems that the most recent 5 commits from main are somehow included in the PR after I rebased. I'm not sure why.

This is because your PR is opened against next-minor branch, and should have been rebased against it. git rebase -i interactive mode is very helpful in such situations, and you can easily remove a few commits.

Ah. I rebased from main but there are commits there that aren't in next-minor yet so they got picked up. I'm not sure what the right thing is here. Should I make a new PR with the pglite stuff on top of next-minor instead of main?

No need to make a new branch. Just make git rebase -i HEAD~12 and inside the interactive editor, delete rows with commits you don't want

Ah got it, fixed

… changes from Effect-TS#5104

bwbuchanan · 2026-02-14T15:16:28Z

Could this use PgJson instead of PgliteJson for the custom sql type, and align the type so that the argument is unknown instead of [unknown] to enable drop-in compatibility with @effect/sql-pg?

Also, it seems that PgArray was recently removed from @effect/sql-pg, so PgliteArray may now also be redundant?

evelant · 2026-03-19T12:08:59Z

I've somewhat put this on hold as I'm not sure if adding a pglite lib to core effect really makes sense. Pglite is a very useful idea but the progress of the project has been glacial and I'm not sure if it is stable and supported enough to warrant a package in effect.

bwbuchanan · 2026-03-19T12:33:38Z

pglite is essential to the test suite of a codebase I maintain, because it would drastically increase the test run time to spin up a "real" Postgres contanier.

I dropped your code from this PR directly into my repo, which made it possible to upgrade effect past the point where pg became the backend for @effect/sql-pg. Claude was then able to one-shot a port to effect v4 beta.

Unfortunately this is far enough outside of my usual responsibilities that I can't allocate any time to helping with this PR right now, but for anyone else in the same situation of needing to test their Postgres-dependent effect code, I would suggest pointing your coding agent at this PR and asking it to sort it for you.

evelant · 2026-03-20T14:40:57Z

Heh of course I made that comment yesterday only to find that electric-sql has just finished their big rework of PGlite. It may be a lot more stable/usable now. I'll try to bring this PR up to speed again soon.

evelant · 2026-05-15T20:20:40Z

It looks like someone took this and patched it up for v4 and it's now merged, thanks @blntrsz! Going to close this since v4 now has @effect/sql-pglite.

github-project-automation Bot added this to PR Backlog Apr 3, 2025

github-project-automation Bot moved this to Discussion Ongoing in PR Backlog Apr 3, 2025

effect-bot force-pushed the next-minor branch 5 times, most recently from 02d328c to 1ea11b9 Compare April 4, 2025 15:48

effect-bot force-pushed the next-minor branch 19 times, most recently from 4cba33b to 5c9e640 Compare April 14, 2025 00:33

effect-bot force-pushed the next-minor branch 2 times, most recently from f265f41 to 0fd314b Compare April 14, 2025 01:59

nikelborm reviewed Jun 4, 2025

View reviewed changes

nikelborm mentioned this pull request Jun 26, 2025

Updated deprecated OTel span attributes names and values #5104

Closed

5 tasks

ooga reviewed Jul 2, 2025

View reviewed changes

Comment thread packages/sql-pglite/src/PgLiteClient.ts Outdated

tim-smart requested changes Jul 23, 2025

View reviewed changes

tim-smart reviewed Jul 23, 2025

View reviewed changes

mikearnaldi and others added 2 commits December 13, 2025 05:00

Add logs to first propagated span (Effect-TS#5710)

05284e2

RcMap: support dynamic idleTimeToLive values per key (Effect-TS#5859)

f752fc3

Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Tim Smart <hello@timsmart.co>

nikelborm reviewed Dec 17, 2025

View reviewed changes

evelant and others added 7 commits December 17, 2025 10:50

Add @effect/sql-pglite package

af159fc

Updated deprecated OTel Resource attributes names and values matching…

b654fe0

… changes from Effect-TS#5104

fixed wrong otel span attributes and fixed option variable name mismatch

38025cc

fix pglite executeValues

575c35b

fix pglite executeValues, listen, run, cleanup test

19e993e

cleanup pglite to align with recent sql changes

400553f

pglite: fix casing of filenames, align scripts

cb5dfd6


		test("should work", () => expect(true))
		describe("PgliteClient", () => {

Uh oh!

Conversation

evelant commented Apr 3, 2025

Type

Description

Related

Uh oh!

changeset-bot Bot commented Apr 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🦋 Changeset detected

Uh oh!

evelant commented Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

evelant commented May 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

evelant commented Jun 3, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

evelant commented Jun 18, 2025

Uh oh!

nikelborm commented Jun 30, 2025

Uh oh!

Uh oh!

nikelborm commented Jul 8, 2025

Uh oh!

evelant commented Jul 22, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bwbuchanan commented Nov 4, 2025

Uh oh!

evelant commented Nov 4, 2025

Uh oh!

thewilkybarkid commented Nov 27, 2025

Uh oh!

evelant commented Dec 17, 2025

Uh oh!

This comment was marked as resolved.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nikelborm Dec 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nikelborm Dec 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bwbuchanan commented Feb 14, 2026

Uh oh!

evelant commented Mar 19, 2026

changeset-bot Bot commented Apr 3, 2025 •

edited

Loading

evelant commented Apr 4, 2025 •

edited

Loading

evelant commented May 1, 2025 •

edited

Loading

nikelborm Dec 17, 2025 •

edited

Loading

nikelborm Dec 17, 2025 •

edited

Loading