There is also Apache Arrow that uses a similar use case. Maybe this is worth considering: https://arrow.apache.org/docs/python/memory.html#on-disk-and...

Flatbuffers can generate the json-schema for you out of the fbs files.

Agree, the ecosystem for flatbuffer and the perfomance is great. Here is the language support: https://google.github.io/flatbuffers/flatbuffers_support.htm...

Totaly agree. I use flatbuffer and it is a lot faster than e.g Json: https://google.github.io/flatbuffers/flatbuffers_benchmarks....

you have to use it since PostgreSQL does not support limit within a with expression.

You can use a with-expression. It's true that you can't use `limit` to limit it, but you can use a `where` condition. This is equivalent to cribwi's example:

  WITH RECURSIVE
    t AS (
      SELECT 0 id
      UNION ALL
      SELECT id + 1
        FROM t
        WHERE id < 1000000
    )
    SELECT id, random() FROM t;

It returns 1000001 rows, but so does cribwi's.

There was a podcast lateley on FLOSS Weekly: https://twit.tv/shows/floss-weekly/episodes/451

Thanks.

ChaiScript has a stackbased memory model so this would fit.

Wren has a garbage collector, so this would be probalby no option for me.

Euphoria is not on github so I have to search first how the gc works. Thanks again