You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Ability to parse strings that look like timestamps to timestamp type. Need to consider whether only timestamp type should be supported as most JSON libraries stick to ISO8601. It might also be inefficient to use regex for timestamps, so the user should provide a hint of which columns to convert to timestamps
The text was updated successfully, but these errors were encountered:
It seems the behavior here is not consistent. TemporalValues can get parsed to u64 if they look like u64 otherwise it is a null, but only while dealing with build_struct_array whereas for build_nested_list_array an error is returned.
It could be done as done in csv_reader with regex and extra strings describing possible formats. Or we could have parser as closures (eg : Fn(String) -> Type)) stored in the reader/decoder(/Schema?)
Note: migrated from original JIRA: https://issues.apache.org/jira/browse/ARROW-4803
Ability to parse strings that look like timestamps to timestamp type. Need to consider whether only timestamp type should be supported as most JSON libraries stick to ISO8601. It might also be inefficient to use regex for timestamps, so the user should provide a hint of which columns to convert to timestamps
The text was updated successfully, but these errors were encountered: