fix: sync diagnostics row count

[stevensJourney]

Overview

The Sync diagnostics app currently displays a row count per table in the sync diagnostics page.

This row count is currently obtained by counting the ops per table in ps_oplog.

SELECT row_type as name, count() as count, sum(length(data)) as size FROM ps_oplog GROUP BY row_type

TLDR: The above approach works if the rows only originate from a single bucket definition in the sync config.

More details:

ps_oplog stores ops with the schema:

CREATE TABLE ps_oplog(
  bucket INTEGER NOT NULL,
  op_id INTEGER NOT NULL,
  row_type TEXT,
  row_id TEXT,
  key TEXT,
  data TEXT,
  hash INTEGER NOT NULL) STRICT;

Where the ops are related to a bucket.

There will be multiple ps_oplog entries for a row if it is referenced from multiple sync bucket definitions. E.g.

bucket_definitions:
  # "global" sync rules will sync to all users. Note that 'global' is not a reserved word - buckets can be named anything
  global:
    data:
      # Sync all rows
      - SELECT * FROM organizations

  by_user:
    parameters:
      - SELECT request.user_id() as user_id
    data:
      - SELECT * FROM organizations where user_id = bucket.user_id

This causes the rowcount to be larger than the actual number of rows (which are deduplicated).

This updates the query logic to query the views directly for a row count.

Implementation

The Tables table now contains a Row Count column which should correlate to the number of synced rows present.

An additional Synced Count column is added to show the previous value as recommended here.

[changeset-bot]

🦋 Changeset detected

Latest commit: 58c1ce4

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@powersync/diagnostics-app	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[rkistner]

Happy with the updates, just added some suggestions for wording