Reading Delta Tables in Fabric

[bill-ash]

Getting the following error:

IOException: IO Error: Parquet file is likely corrupted, cannot have dictionary offsets without seeing a non-empty dictionary first.

when reading a delta table in Fabric. No issues when reading the same table using spark.

When I choose a subset of columns the query will complete successfully ie:

# throws an error
duckdb.sql("select * from delta_scan('/lakehouse/default/Tables/<schema>/<table_name>'")

# completes without error 
duckdb.sql("select col1, col2, from delta_scan('/lakehouse/default/Tables/<schema>/<table_name>'")

Will try to nail down the exact columns and their types that are causing issues but maybe someone else recognizes the error. Thanks!

[samansmink]

This should be fixed through duckdb/duckdb#14268. Would you mind trying again on duckdb v1.1.3?