You are not logged in.
I'm in trouble with creating a file delimited metadata from a *csv file.
An other type than "string" causes an error message.
Maybe a wrong field or row separator causes the problem. Can you help me to find out where the mistake is?
The *csv file has the following schema:
Last edited by Piero001 (2012-05-21 12:05:32)
I suggest you read the value as a String, just because it is a String as the quotes tell us. Than, afterwards you can use a tMap or nother component to convert the type of the variable from String to Integer.
Hope this helps,
thanks for your answer.
Im confused with the fact that talend automaticly suggests Integer for the "Transaktionsummer" Column as you can see in screenshot 4, but when I run the job it says Error for string: " "123123" "
why can't Talend read the numbers without the double quotation? I think it uses the wrong field seperator.
I don't know how to convert columns with tMap into the right typ. Is there no other solution for the problem, like changing to the right corresponding charakter? (Screenshot 3)
thanks for your help,
I'm sorry, I didn't look at your screens as good as I should have.
You are absolutly right: Talend can read your data without you needing to convert the type of the field. You should just check the CSV Options checkbox (or the CSV radio button under the Escape Char Settings in the wizard).
This will make Talend read the integer without quotes and make them valid.
Hope this helps.
thank you again! That's exactly what I'm looking for.
But your tipp causes some new troubles: Changing the Escape Char Settings to CSV raises the number of Columns up to 255 (!) - now Talend even involved the empty cells of the *csv file.
Second Question: Which Escape Char and which Text Enclosure should i choose for my Data? I can only choose between
but shouldn't the right escape char for my file type not looking like this """" ?
thanks for all your help,
The one you should use is "\"" This is an escaped version of " (the double quote)
Assuming you choose this for text enclosure and you're still getting this much columns you might want to check your input file. Does this one (when opening in excel for example) show this much rows? I once saw someone who accidentally put some formatting on column XE, this cause his CSV export to export all columns up until col XE...
Hope this helps!
Thanks to you i can now choose every typ I want: Integer, Char.
The new problem is definitly caused by Talend! I created the following example file with MS notepad and tried to generate metadata in Talend.
You can try it by yourself, copy and paste the following lines to the MS Notepad and save it as an .csv file
If you generate metadata from that source, checking the CSV box in step 3 and click the refresh button, the columns will raise up to 255...
Here are my current settings:
";" = Field Seperator
"\n" =Row Seperator
"\"" = Escape Char
"\"" = Text Enclosur
Does someone know how to deal with that???
Last edited by Piero001 (2012-05-21 18:01:24)
I've tried your suggested csv file, but couldn't reproduce your column issue. This might be due to the fact I'm using another version of Talend. Currently I'm on 4.2.3. What's yours?
Good morning to you too Piero,
Well, looks like we've actually found a bug here if it turns out 5.1.0 generates more columns than needed.
I might try a 5.1.0 version here too today if I can find the time. Maybe on some other platform as well. I'm developing on a Windows 7 64 just like you, but I have some Linux at hand on a VM.
Maybe Pedro or Shong can shine a light on this as well. Can't you guys?
I'll keep you all posted.
I finally found out that the column problem only appears in the preview window. Everything works fine if you just ignore the preview window not to much columns when executing a job.
think it is a bug in version 5.1.0, it's not a big deal but it's just curious.
thanks for all your help!!