What causes DELTA_STATS_COLLECTION_FAILED?

A column type (e.g., STRUCT, MAP, ARRAY, or a very long STRING) does not support file statistics collection. The table has more than 32 columns and Delta only collects statistics for the first 32 by default. Autoloader or a streaming job wrote rows that caused a stats computation error on a nested schema. Custom data types or UDTs used in the schema are not supported by Delta's stats collection engine

How do I fix DELTA_STATS_COLLECTION_FAILED?

For unsupported column types, exclude them from stats collection: `ALTER TABLE t SET TBLPROPERTIES ('delta.dataSkippingNumIndexedCols' = )` where n is the count of columns to include.. If the failing column is a complex type (STRUCT/MAP/ARRAY), set `delta.dataSkippingNumIndexedCols` to exclude it from the stats window.. Check whether the write completed successfully despite the stats failure — data is written even when stats collection fails.. Run `ANALYZE TABLE t COMPUTE STATISTICS` to rebuild stats after addressing the root cause.

Low severitysql

Power BI Refresh Error:
DELTA_STATS_COLLECTION_FAILED

What does this error mean?

Delta Lake failed to collect file-level statistics (min/max values, null counts) for one or more columns during a write operation. Statistics are used for data skipping; their absence degrades query performance but does not corrupt data.

Common causes

1A column type (e.g., STRUCT, MAP, ARRAY, or a very long STRING) does not support file statistics collection
2The table has more than 32 columns and Delta only collects statistics for the first 32 by default
3Autoloader or a streaming job wrote rows that caused a stats computation error on a nested schema
4Custom data types or UDTs used in the schema are not supported by Delta's stats collection engine

How to fix it

1For unsupported column types, exclude them from stats collection: `ALTER TABLE t SET TBLPROPERTIES ('delta.dataSkippingNumIndexedCols' = <n>)` where n is the count of columns to include.
2If the failing column is a complex type (STRUCT/MAP/ARRAY), set `delta.dataSkippingNumIndexedCols` to exclude it from the stats window.
3Check whether the write completed successfully despite the stats failure — data is written even when stats collection fails.
4Run `ANALYZE TABLE t COMPUTE STATISTICS` to rebuild stats after addressing the root cause.

Frequently asked questions

Is my data lost when stats collection fails?

No — stats are metadata used for query optimization, not the data itself. The write succeeds and the data is readable. Only query performance is affected.