As discussed on dev@: https://lists.apache.org/thread/onzgogx2c2djxs0wbhmvqp2dbx7kjf6o "[ARROW-17255] Logical JSON type in Arrow"
It would be interesting to have JSON parsing/serializing compute functions that operate on columns of (stringified) JSON records. For parsing, the problem is we need to know the output schema without being able to look at the data, so we would probably only be able to decode into a map[string, union] type at best. And/or we could offer "extraction" functions akin to what things like SQLite and Postgres provide (at the cost of having to reparse the JSON over and over).
Also see ARROW-17255 for a logical JSON type.
Reporter: David Li / @lidavidm
Related issues:
Note: This issue was originally created as ARROW-17268. Please see the migration documentation for further details.
As discussed on dev@: https://lists.apache.org/thread/onzgogx2c2djxs0wbhmvqp2dbx7kjf6o "[ARROW-17255] Logical JSON type in Arrow"
It would be interesting to have JSON parsing/serializing compute functions that operate on columns of (stringified) JSON records. For parsing, the problem is we need to know the output schema without being able to look at the data, so we would probably only be able to decode into a
map[string, union]type at best. And/or we could offer "extraction" functions akin to what things like SQLite and Postgres provide (at the cost of having to reparse the JSON over and over).Also see ARROW-17255 for a logical JSON type.
Reporter: David Li / @lidavidm
Related issues:
Note: This issue was originally created as ARROW-17268. Please see the migration documentation for further details.