question "Error parsing row: file" Table consists of multiple csv files in Support Questions

question "Error parsing row: file" Table consists of multiple csv files in Support Questions https://community.cloudera.com/t5/Support-Questions/quot-Error-parsing-row-file-quot-Table-consists-of-multiple/m-p/286571#M212523 Dear community,   I have created a new datatable by uploading a csv file (incl. header / the csv file contains data about a specific months) to HDFS (via Hue). Afterwards I have cleared the cache and uploaded the other csv files (all following csv files have the same column order BUT NO HEADER; average size of every monthly csv file: ~2-4 GB; number of columns: 54). Typical procedure after uploading a new csv file to the database: INVALIDATE METADATA database_xy   When I send a query where every column shall be displayed I get the following Error Messages in Impala: Error converting column: 6 to TIMESTAMP Error converting column: 8 to TIMESTAMP Error converting column: 23 to TIMESTAMP Error converting column: 50 to TIMESTAMP Error converting column: 35 to TIMESTAMP Error converting column: 43 to TIMESTAMP   Information for this columns are available after 4 months. Till then there are only NULL values.   Query to reproduce these error messages:    SELECT * FROM database_xy LIMIT 100   For a specific TIMESTAMP column: SELECT min(exp_date) FROM database_xy   Error (Just a sample of the log box in Hue):  Error parsing row: file: hdfs://blabla/foo_042019.csv, before offset: 2432696320 Error converting column: 21 to TIMESTAMP Error parsing row: file: hdfs://blabla/foo_032019.csv, before offset: 1895825408 Error converting column: 21 to TIMESTAMP Error converting column: 21 to TIMESTAMP Error parsing row: file: hdfs://blabla/foo_022019.csv, before offset: 2969567232 Error converting column: 21 to TIMESTAMP Error converting column: 21 to TIMESTAMP     When I run the queries in Hive I get no error messages at all. How come? And how do i get rid of those error messages in Impala?     Information about how I created the csv-files locally: First CSV: Python (Pandas): Set Options: Separator: Pipe, (only for first csv:) header=True, index=False (so there is no additional useless index column)   Subsequent CSVs: Python (Pandas): Set Options: Separator: Pipe, header=False, index=False (so there is no additional useless index column)   When I created the table with the first CSV in Hue I selected the following options: Field Separator: Pipe Record Separator: New line Quote Character: Double Quote   Afterwards I have uploaded all the other CSVs in the database's folder to add the new months and invalidated the metadata.   Thank you for your help in advance! I hope you enjoyed the Christmas holidays and I wish you a happy New Year's Eve!   Best somedatadude Mon, 30 Dec 2019 15:08:59 GMT somedatadude 2019-12-30T15:08:59Z