4 ms·

PostgREST 11 is in pre-release too - Steve (the maintainer) wrote about some of the upcoming features here: https://supabase.com/blog/postgrest-11-prerelease h

PostgREST 11 is in pre-release too - Steve (the maintainer) wrote about some of the upcoming features here: https://supabase.com/blog/postgrest-11-prerelease https://supabase.com/blog/postgrest-11-prerelease (Disclosure: this points to the supabase blog and I work at supabase (as does Steve)) Some of the features include - Spreading related tables - similar to a JS "..." spread operator to flatten a response - Order by on related tables - "Anti-Joins" - e.g. filter the rows where the related table is null.

MuffinFlavored 4y ago
I'm kind of confused why GROUP BY and "DISTINCT" aren't prioritized as features. I know the workaround (create a view) but I'm a little curious why it isn't seen as a "I expected this to already be there" shortcoming for these rather "basic" features. I'm well aware the open source attitude means "go and fork it + PR it yourself if you want it so bad". Easier said than done as an outsider into a massive Haskell codebase. Transactions (2015): https://github.com/PostgREST/postgrest/issues/286 https://github.com/PostgREST/postgrest/issues/286 GROUP BY/DISTINCT (2017): https://github.com/PostgREST/postgrest/issues/915 https://github.com/PostgREST/postgrest/issues/915
- SOLAR_FIELDS 4y ago
  I agree 100%, I have been using supabase a lot over the last few months and the #1 glaring wart on an otherwise pretty good system is the lack of aggregates. Feels like a table stakes feature to at least support the very basic stuff like max, average, sum since any app that isn’t the most basic of basic apps is going to need those. And yes, I’m aware you can use an RPC or a view to work around that, but doing so very much black boxes your application logic into SQL and should be heavily avoided for several reasons and using those things come with their own limitations as well (can’t use Supabase Realtime on RPC or view, as one example).

MuffinFlavored 4y ago

I'm kind of confused why GROUP BY and "DISTINCT" aren't prioritized as features. I know the workaround (create a view) but I'm a little curious why it isn't seen as a "I expected this to already be there" shortcoming for these rather "basic" features. I'm well aware the open source attitude means "go and fork it + PR it yourself if you want it so bad". Easier said than done as an outsider into a massive Haskell codebase. Transactions (2015): https://github.com/PostgREST/postgrest/issues/286 https://github.com/PostgREST/postgrest/issues/286 GROUP BY/DISTINCT (2017): https://github.com/PostgREST/postgrest/issues/915 https://github.com/PostgREST/postgrest/issues/915

SOLAR_FIELDS 4y ago
I agree 100%, I have been using supabase a lot over the last few months and the #1 glaring wart on an otherwise pretty good system is the lack of aggregates. Feels like a table stakes feature to at least support the very basic stuff like max, average, sum since any app that isn’t the most basic of basic apps is going to need those. And yes, I’m aware you can use an RPC or a view to work around that, but doing so very much black boxes your application logic into SQL and should be heavily avoided for several reasons and using those things come with their own limitations as well (can’t use Supabase Realtime on RPC or view, as one example).