Hi, I did search for similar issues but didn't find a perfect match for my case:     CREATE EXTERNAL TABLE processed_data LIKE PARQUET '/user/myuser/spark/proyecto1/date=2015-08-17/hour=21/part-00087-7a6eac40-0db4-4368-894c-72588f80dfd6.c000.snappy.parquet' STORED AS PARQUET LOCATION '/user/myuser/spark/proyecto1'   The table creation works, but any query to the table will return 0 rows.   Query: SHOW CREATE TABLE processed_data +---------------------------------------------------------------------------------------------------------------------------+ | result | +---------------------------------------------------------------------------------------------------------------------------+ | CREATE EXTERNAL TABLE josemidb.processed_data ( | | window STRUCT<start:TIMESTAMP,end:TIMESTAMP> COMMENT 'Inferred from Parquet file.', | | terminating_cellid BIGINT COMMENT 'Inferred from Parquet file.', | | termination_id BIGINT COMMENT 'Inferred from Parquet file.', | | termination_reason BIGINT COMMENT 'Inferred from Parquet file.', | | numcalls BIGINT COMMENT 'Inferred from Parquet file.' | | ) | | STORED AS PARQUET | | LOCATION 'hdfs://myhost:8020/user/myuser/spark/proyecto1' | | TBLPROPERTIES ('COLUMN_STATS_ACCURATE'='false', 'numFiles'='18', 'numRows'='-1', 'rawDataSize'='-1', 'totalSize'='67251') | +---------------------------------------------------------------------------------------------------------------------------+ This shows numFiles = 18 which is right, but numRows = -1 😞   I think my problem is that I'm not handling partitions the right way. Can you help me with this?   Thank you. ... View more