So essentially you have three options: a) you know that you only have data from that day in the temp table. So you can just drop the column completely and do a static partition load specifiyng the date in the format you want. Essentially just delete that one column during the insert. INSERT OVERWRITE TABLE final partition ( daily_date=20160412 ) select id, name, <all columns but daily_date> from staging; b) you have one or two days in there ( mid day ) so you just filter on the daily_date column but do the same thing otherwise ( in this case you dont want overwrite since the edge days will be loaded twice INSERT APPEND TABLE final partition ( daily_date=20160412 ) select id, name, <all columns but daily_date> from staging where daily_date="12/04/2016"; INSERT APPEND TABLE final partition ( daily_date=20160411 ) select id, name, <all columns but daily_date> from staging where daily_date="11/04/2016"; c) you use dynamic partitioning and let hive figure it out INSERT APPEND TABLE final partition ( daily_date ) select id, name, , all columns but daily_date, regex... as daily_date from staging; The only thing is that your partition column needs to be last ( also I like ints instead of string i.e. 20160412 as an integer makes for easier computations like day + 7 or something ) You also need to enable some parameters for dynamic partitioning: http://www.slideshare.net/BenjaminLeonhardi/hive-loading-data
... View more
