We have a fixed a bug that caused Int32 and Int64 values to sync incorrectly from Parquet files. Previously, our syncs did not include the scale of these values (for example, a value of
85.00 was loaded into the destination as
8500000). We now scale these values correctly. If any of your Int32 or Int64 values were incorrectly synced prior to this fix, you will need to initiate a re-sync to correct them.
We now support syncing headerless delimited format files (CSV, TSV, log) for the following file connectors:
We create generic column names for CSV files without a header line. Learn more in our Files connector documentation. This feature is in beta and available to all customers.
We now support
jsonl file extensions in all file connectors.
The following file connectors can now sync Parquet files:
We support Parquet format 2.4.0. This feature is in BETA.
Fivetran's incremental file update mechanism now re-syncs files that were created at the same time as the last observed cursor position.
We now support Avro files in all file replication connectors.
We now parse ISO DATE format (yyyy-mm-dd-hh:mm:ss) correctly.
If your file store contains an invalid .zip file, we will skip syncing it and issue a warning.