Meilisync guides: Add note regarding primary key field
guimachiavelli opened this issue · comments
A user mentioned they ran into issues with using meilisync
when their dataset contains several fields ending in _id
:
[…] I had to add the pk (primary key) attribute to the tables in the config.yml because my tables have more than one column ending in 'id'. Note that the pk YAML attribute is not mentioned in the documentation, but the reported error suggests its existence.
Pinging @meilisearch/devrel-team, the official maintainers of the meilisync
guides.
Tasks
- verify with SME (@alallema) whether the info checks out and if this scenario also applies to PostgreSQL
- update
/learn/primary_db_sync/meilisync_mysql.mdx
- if necessary, update
/learn/primary_db_sync/meilisync_postgre.mdx
References
- User report on Discord: https://discord.com/channels/1006923006964154428/1156967511313621022/1166073598579048559
- SME: @alallema
Indeed, if you have several fields with id
in them, you'll get an index_primary_key_multiple_candidates_found
error, at which point meilisync proposes a way to set the primary key field as it is:
sync:
--
- table: POSTGRES_TABLE_NAME_1
index: MEILISEARCH_INDEX_NAME_1
pk: _id
full: false
I just checked the meilisync code because the user is right this is not in the documentation