Currently, the querying of Parquet files is not handled correctly. Columns are interpreted correctly, just not the tables part (as expected).
Following an example of a Parquet query, using OPENROWSET:
SELECT q.MY_FIELD01
FROM OPENROWSET (BULK '/DATA_STORE1/1/data/date=2022-12-25/*'
,DATA_SOURCE = 'parquet-datasource'
,FORMAT = 'PARQUET') q
WHERE q.MY_FIELD02 >= 'value';
The current output is as follows:
| Key |
Value(s) |
| Table names |
OPENROWSET |
| Table names |
BULK |
| Table names |
DATA_SOURCE |
| Table names |
FORMAT |
| Column names |
q.MY_FIELD01 |
| Column names |
q.MY_FIELD02 |
But, as you might have guessed, OPENROWSET, BULK, DATA_SOURCE and FORMAT are actually all keywords.
For Table names the correct output should have been /DATA_STORE1/1/data/date=2022-12-25/*
Currently, the querying of Parquet files is not handled correctly. Columns are interpreted correctly, just not the tables part (as expected).
Following an example of a Parquet query, using
OPENROWSET:The current output is as follows:
But, as you might have guessed,
OPENROWSET,BULK,DATA_SOURCEandFORMATare actually all keywords.For Table names the correct output should have been
/DATA_STORE1/1/data/date=2022-12-25/*