The following error message is displayed when the LOAD DATA command is executed by a Spark SQL job to import data to a DLI table:
error.DLI.0001: IllegalArgumentException: Buffer size too small. size = 262144 needed = 2272881
error.DLI.0999: InvalidProtocolBufferException: EOF in compressed stream footer position: 3 length: 479 range: 0 offset: 3 limit: 479 range 0 = 0 to 479 while trying to read 143805 bytes
The data volume of the file to be imported is large and the value of spark.sql.shuffle.partitions is too large. As a result, the cache size is insufficient.
Decrease the spark.sql.shuffle.partitions value. To set this parameter, perform the following steps: