Search code examples
hadoophivehiveqlhive-serdehiveddl

Auto selected delimiter while using csv-serde


Have created a table using:

 )   
ROW FORMAT SERDE 'com.bizo.hive.serde.csv.CSVSerde'
with serdeproperties(
"field.delim"="|"
)   

When I tried to upload a file, that is comma-delimited, it has been uploaded correctly. How is it possible, if I had configured pipe as a delimiter for the data? Does csv-Serde has some "smart" way of identifying the delimiter?


Solution

  • As per documentation valid CSVSerde properties are

       "separatorChar" = "\t",
       "quoteChar"     = "'",
       "escapeChar"    = "\\"
    

    Default separator, quote, and escape characters if unspecified

    DEFAULT_ESCAPE_CHARACTER \
    DEFAULT_QUOTE_CHARACTER  "
    DEFAULT_SEPARATOR        ,
    

    field.delim is not supported property which is ignored and it uses default setting. This property is valid for another SerDe MultiDelimitSerDe