Impala supports Read Optimized Queries on Hudi tables as an experimental feature. Read Optimized Queries treat the latest table snapshot as the commit or compaction action, expose only base/columnar files in latest file slices, and guarantee the same columnar query performance as a non-Hudi columnar table.
File Type | Format | Compression Codecs | Impala Can CREATE? | Impala Can INSERT? |
---|---|---|---|---|
Hudi | Structured | Snappy, gzip, zstd, lz4; currently Snappy by default | Yes, support for Read Optimized Queries is experimental. | No. Create an external table in Impala. Set the table location to the Hudi table directory. Alternatively, create the Hudi table in Hive. |