Hello!
I am having difficulties reading parquet Timestamp into a Synapse SQL.
I am using a straightforward query:
SELECT
BIRTHDAY
FROM
OPENROWSET(
BULK 'https://{account}.dfs.core.windows.net/root/parquet/part-00000-tid-7858592301442936858-0ac91229-4635-4c53-92ed-a5e1be35e33c-0-1-c000.snappy.parquet',
FORMAT='PARQUET'
) AS [result]
And getting the following error:
Error handling external file: 'Inserting value to batch for column type DATETIME2 failed. Invalid argument provided.'. File: 'https://{account}.dfs.core.windows.net/root/parquet/part-00000-tid-7858592301442936858-0ac91229-4635-4c53-92ed-a5e1be35e33c-0-1-c000.snappy.parquet'.
I have checked the data and I didn't see any problems, some records do have Timestamp as of '0001-01-01 00:00:00' but I consider this is acceptable for datetime2 data type.
I am able to read the file without this column or with top 100 including birthday, so it is more or less about the data I think.
Can you provide any suggestions or tips on this question?
Thanks!
EDIT:
The problem persist with some other operations (as of COPY bulk from parquet/ORC to internal table).
following datetime is an issue and raises error:
0001-01-01 02:00:00
Please check on your side




?