This section provides information on improvements to the
|D s item|
Mismatched values are no longer published as null values in CSV outputs
In prior releases, when a file was published in CSV format, any values that were mismatches for a column's data type were written as null values, which could lead to loss of data that was meaningful to downstream systems.
Beginning in this release, mismatched values are written out in CSV format as String values by default.
NOTE: The ability to write out mismatched values in CSV outputs is enabled by default in new flows and CSV publishing actions. For existing CSV outputs, the prior behavior is maintained.
NOTE: This capability applies only to CSV outputs at this time. In the future, it will be applied to other non-schematized outputs, such as JSON.
Tip: When visual profiling is enabled, you can still identify the values in the generated results that are mismatched for their column data types.
As needed, you can configure the ability to write out mismatches for CSV outputs for individual publishing actions. For more information, see File Settings.
Snowflake date publishing improvements
In prior releases, the
|D s webapp|
Beginning in this release, the
|D s webapp|
For more information, see Snowflake Data Type Conversions.
Data type inference and row split inference run on more data