The pgec's discuss from shortishly

pgec's Issues

Feature idea: PostgreSQL API

I am very impressed by the idea behind this product.

How realistic is the idea of implementing a PostgreSQL API?

The PostgreSQL API should behave like a real PostgreSQL Server.
Read operations are served from the in memory cache.
Write operations are proxyed to the PostgreSQL server.
Like the current REDIS API.

The biggest advantage of a PostgreSQL API would be the integration of PGEC without code changes for applications.

PostGIS support

Hi @shortishly,

I learned about this project from Scaling Postgres#254 - very interesting work!

I tried to replace the test data with my own data and encountered an issue. I'm using PostGIS (geometry and raster) data, which basically worked in the REST API, but unfortunately was encoded binary in the output:

Maybe this is how pgec handles all non-simple UDTs.

It would be great to add ST_AsBinary(), ST_AsText() or ST_AsGEOJSON() for the output PostGIS vector geometries of pgec (REST, Memcached and Redis API) and maybe ST_AsBinary() for PostGIS raster objects.

I'm happy to contribute if you point me to the code - although I don't have any knowledge in Erlang.

Best regards,
Johannes

JSONB Support

Hi Peter,

I encountered an issue when dealing with JSONB columns in a publication. When I add the test data like this:

CREATE TABLE public.json_sample (id serial PRIMARY KEY, keyval jsonb);
INSERT INTO public.json_sample (keyval) VALUES ('{"foo": "bar", "baz": 1}');

When I change the publication to this dataset, pgec seem to crash and the REST API is not available anymore:

=CRASH REPORT==== 3-Mar-2023::19:10:32.062821 === <0.6203.0> proc_lib:crash_report/4:539
  crasher:
    initial call: cowboy_stream_h:request_process/3
    pid: <0.6203.0>
    registered_name: []
    exception error: bad argument
      in function  ets:safe_fixtable/2
         called as ets:safe_fixtable(osm_rivers_sample,true)
         *** argument 1: the table identifier does not refer to an existing ETS table
      in call from ets:foldl/3 (ets.erl, line 632)
      in call from pgec_h:init/3 (src/pgec_h.erl, line 51)
      in call from cowboy_handler:execute/2 (src/cowboy_handler.erl, line 37)
      in call from cowboy_stream_h:execute/3 (src/cowboy_stream_h.erl, line 306)
      in call from cowboy_stream_h:request_process/3 (src/cowboy_stream_h.erl, line 295)
    ancestors: [<0.6201.0>,<0.217.0>,<0.216.0>,ranch_sup,<0.116.0>]
    message_queue_len: 0
    messages: []
    links: [<0.6201.0>]
    dictionary: []
    trap_exit: false
    status: running
    heap_size: 376
    stack_size: 28
    reductions: 323
  neighbours:

=ERROR REPORT==== 3-Mar-2023::19:10:32.064281 === <0.6201.0>
Ranch listener pgec, connection process <0.6201.0>, stream 1 had its request process <0.6203.0> exit with reason badarg and stacktrace [{ets,safe_fixtable,[osm_rivers_sample,true],[{error_info,#{cause => id,module => erl_stdlib_errors}}]},{ets,foldl,3,[{file,"ets.erl"},{line,632}]},{pgec_h,init,3,[{file,"src/pgec_h.erl"},{line,51}]},{cowboy_handler,execute,2,[{file,"src/cowboy_handler.erl"},{line,37}]},{cowboy_stream_h,execute,3,[{file,"src/cowboy_stream_h.erl"},{line,306}]},{cowboy_stream_h,request_process,3,[{file,"src/cowboy_stream_h.erl"},{line,295}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,240}]}]

Could look into this?

shortishly / pgec Goto Github PK

pgec's Issues

Feature idea: PostgreSQL API

PostGIS support

JSONB Support

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent