Date alteration during tMap

[Anonymous]

Hi All, 
I am migrating data for a client from .csv to Salesforce. The issue that I am having is that for a large number of rows, all the dates are decreasing by one month.
The date values are being converted from a string to a date (automatically, not using an expression). I have pulled the data from the tmap into a csv and I believe it is happening during that part of the transformation process. If it were the day value that was changing, I would potentially attribute the issue to salesforce locale or timezone issues. The fact that is the month that is decreasing by a value of one is strange.
An example would be:
Source Data

20160212

20160215

Output

12/01/2016

15/01/2016

Could this just be an issue in the source data? Hopefully someone can help!

Thanks in advance,

Chris

[TRF]

Hi,
Strange result...
I use to deal with Salesforce and Talend and never see something like that. 
However, what are your locale and timezone on Salesforce side?

Regards,
TRF

[Anonymous]

Hi,
Yeah it's very peculiar. It is also not happening for all rows. The field is also of type date, not datetime.
All Users Locale is United Kingdom, but the timezone varies as it's a global org. I am on GMT +2.

[TRF]

I think the problem come from your date representation. 
Do you specify a TZ at upsert or are the dates aligned on GMT?
Check the values received by Salesforce (Quick Find box, then select Bulk Data Load Jobs).

[Anonymous]

Hi TRF, 
I had a look at the request and it is being passed an incorrect date. Talend is set at yyyy-mm-dd. The source value is a just a string, e.g. 20160216. I altered the job to output the tmap into a csv, where the value would be 16-01-2016.