With a tMap i am joining a csv file (row4) and a lookup csv file (row3) into a MySQL table. They have a key id for the inner join. Having row4.DATA_INVIO= 2013-04-01 12:47:49 type Date "yyyy-MM-dd HH:mm:ss" row3.DATA_CREAZIONE= 2013-04-01 12:30:57 type Date "yyyy-MM-dd HH:mm:ss" in the Var box of the tMap I calculate the difference in seconds between two dates: TalendDate.diffDate(row4.DATA_INVIO, row3.DATA_CREAZIONE, "ss") Result= -42188 WRONG! But if in the Var box of the tMap I put this (same dates)... TalendDate.diffDate(TalendDate.parseDate("yyyy-MM-dd HH:mm:ss","2013-04-01 12:47:49"), TalendDate.parseDate("yyyy-MM-dd HH:mm:ss","2013-04-01 12:30:57"), "ss") Result = 1012 GOOD! What is happening???? I'm having the same problems with many raws, it's not random, it seems that some rows always give wrong results... Any suggestion is highly appreciated. Stefano
could it be that you had dates in your dataset as Strings - when you had value -42188 and then by CASTing them into Date - you have the correct value?
ok, I have defined the text delimited metadata date field from Date to String row4.DATA_INVIO type String row3.DATA_CREAZIONE type String then I have put this formula: TalendDate.diffDate(TalendDate.parseDate("yyyy-MM-dd HH:mm:ss",row4.DATA_INVIO), TalendDate.parseDate("yyyy-MM-dd HH:mm:ss",row3.DATA_CREAZIONE), "ss") and it works well now. The question remains: why it doesn't work with Date datatype? Thank you very much for the advice. Stefano
System.out.println("value " +value);
Ex: System.out.println("value " +TalendDate.diffDate(TalendDate.parseDate("yyyy-MM-dd HH:mm:ss",row4.DATA_INVIO), TalendDate.parseDate("yyyy-MM-dd HH:mm:ss",row3.DATA_CREAZIONE), "ss"));
use the above syntax in tjava/tjavarow/tjavaflex
For more help on I\O Streams in java please refer the link:
http://way2java.com/io/system-out-println/
