Anyway to get around the filename timestamp requirement for FTP import?
We have a requirement to store the filename of the file we upload to an FTP site on a record in NetSuite. I just spent about 2 hours adding a script and refactoring a flow to use a pre-generated filename and then I ran into this!
That was entirely unexpected and really blows up my plans. Anyone know of a way to get around this "requirement" to have a timestamp in the filename?
If anyone from Celigo sees this: is there any hope of this requirement being removed? By forcing the filename to have a timestamp defined in the File Name field it's impossible to use the filename anywhere else (like I want to). What if an integration partner that uses FTP would not accept files with names that didn't match their specified format? Is there a workaround?
Thank you,
Steve
Comments
Hi Steve,
This requirement is for when you have the Advanced "Skip Aggregation" option in the import step set to true (checked). When Skipping aggregation, this means that multiple files will be uploaded to the FTP endpoint. The Timestamp option is then required because the timestamp (so long as it includes seconds or milliseconds) in the filename will be different for each file that gets created by the import. Otherwise, the files that are being created on the FTP server will be overwritten one after another until the last file is written, causing a loss of data.
To remove this requirement, uncheck the "Skip Aggregation" option in the FTP import step.
Best Regards,
Dan
We also ran into this issue, had to supply a predefined filename per file, and using pagesize=1 had already ensured filenames would always be unique. My workaround was to use a 'constant' timestamp:
which nicely renders out to uniquename.xml
Hi @...,
I missed the replies from you and Bas, I must have ignored the emails. Either way, I came here to see if anyone has updated this and I'm glad to see there are some answers.
Thank you for your response. I think this feature can be improved to provide more flexibility to the developers. Dictating the format of the filename to avoid overwriting is getting in the way. In my case I want to generate a filename that has the order number in it which would take care of the overwriting files concern.
If IO still wanted to help users avoid overwrite bugs the UI could warn users based on the filename value:
Or some such thing. I really hope you'll consider removing this demand that we use a timestamp in the filename.
Thanks again for your reply.
Bas van Ditzhuijzen
Excellent workaround!
I was thinking about using a pagesize of 1 and turning off aggregation, but from the field help in IIO it sounds like it will aggregate ALL the pages.
I'm curious how your filename example has the syntax
Mine is:
The rows part I understand, but I'm curious how you start with data and I start with batch_of_records.
Thank you for your response. I'll try your workaround now.
Please sign in to leave a comment.