Hi,
I took a look to that solution and it cannot be applied for me.
I'm a newbie in Talend and for me is like a moving sand from many points of view.
1. If I make some changes into a tMap component I cannot be sure that partially an older version is still considered. For instance if I remove a mapping when I run the job it still consider the older settings. So, for each change I make I have to remove my components and to add them again, to re-set...
2. I'm working with delimited files (CSV as input) and Oracle as output. Something that pass the compilation test it is possible to not pass the execution one. For instance, if I work with dates from CSV during the compilation some fields are considered dates, but when I try to run the job they are considered strings. If I'm treating them as strings then the compilation step fails.
3. There are a lot of incompatibilities between types: int, float, short, double, ... evaluating strings with ' ' or " " delimiters, and regarding null values treatment. Due to points 1 and 2 is difficult for me to use the accurate functions. Could you give me some basic guidelines?
I give you a concrete example, starting from my CSV file:
SALE_DATE,COUNTRY_ID,PRODUCT_ID,QUANTITY,DISCOUNT
9/21/2019 8:18,1009,114,5,0.12
I have troubles with SALE_DATE and DISCOUNT. In tMap they are Date and Short/Double in both sides.
In my opinion the date follows "M/d/yyyy H:mm" pattern in CSV. How can I simply move the SALES_DATE from CSV in Oracle as date?! I did use TalendDate.parseDate, TalendDate.formatDate and their combination for SALES_DATE, but the result was unparseable or empty value. For DISCOUNT it was similar. In the compilation phase they are Date/Number, but in execution they are Strings. Should I consider them as being strings in CSV?
Thanks
