Parse CSV error

[gaborszollosy]

Hello there!

I'd like to ask if any of you have any experience where you work with a CSV file and you get content like this:

["��i\u0000d\u0000\t\u0000c\u0000r\u0000e\u0000a\u0000t\u0000e\u0000d\u0000_\u0000t\u0000i\u0000m\u0000e\u0000\t\u0000a\u0000d\u0000_\u0000i\u0000d\u0000\t\u0000a\u0000d\u0000_\u0000n\u0000a\u0000m\u0000e\u0000\t\u0000a\u0000d\u0000s\u0000e\u0000t\u0000_\u0000i\u0000d\u0000\t\u0000a\u0000d\u0000s\u0000e\u0000t\u0000_\u0000n\u0000a\u0000m\u0000e\u0000\t\u0000c\u0000a\u0000m\u0000p\u0000a\u0000i\u0000g\u0000n\u0000_\u0000i\u0000d\u0000\t\u0000c\u0000a\u0000m\u0000p\u0000a\u0000i\u0000g\u0000n\u0000_\u0000n\u0000a\u0000m\u0000e\u0000\t\u0000f\u0000o\u0000r\u0000m\u0000_\u0000i\u0000d\u0000\t\u0000f\u0000o\u0000r\u0000m\u0000_\u0000n\u0000a\u0000m\u0000e\u0000\t\u0000i\u0000s\u0000_\u0000o\u0000r\u0000g\u0000a\u0000n\u0000i\u0000c\u0000\t\u0000p\u0000l\u0000a\u0000t\u0000f\u0000o\u0000r\u0000m\u0000\t\u0000m\u0000e\u0000l\u0000y\u0000i\u0000k\u0000_\u0000p\u0000o\u0000z\u0000�\u0000c\u0000i\u0000�\u0000r\u0000a\u0000_\u0000s\u0000z\u0000e\u0000r\u0000e\u0000t\u0000n\u0000e\u0000_\u0000j\u0000e\u0000l\u0000e\u0000n\u0000t\u0000k\u0000e\u0000z\u0000n\u0000i\u0000?\u0000\t\u0000f\u0000u\u0000l\u0000l\u0000_\u0000n\u0000a\u0000m\u0000e\u0000\t\u0000e\u0000m\u0000a\u0000i\u0000l\u0000\t\u0000p\u0000h\u0000o\u0000n\u0000e\u0000_\u0000n\u0000u\u0000m\u0000b\u0000e\u0000r\u0000","\u0000l\u0000:\u00004\u00001\u00005\u00008\u00004\u00002\u00001\u00009\u00003\u00007\u00006\u00001\u00007\u00004\u00001\u00002\u0000\t\u00002\u00000\u00002\u00001\u0000-\u00001\u00000\u0000-\u00001\u00006\u0000T\u00000\u00009\u0000

I get \u0000 before every letter, and all Hungarian characters like (áéíóöőúüű) are changed to a � symbol.

I'd like to use the compose actions to convert the data to json like in this video: https://www.youtube.com/watch?v=sXdeg_6Lr3o

But I can't get past this problem. Even replace doesn't work on removing the '\u0000' parts.

This is how my flow looks. I just want the data from the CSV to appear normally, without the � and the '\u0000' parts, and I don't know what I'm doing wrong.

I appreciate any help!

[tom_riha · Accepted solution]

Hello @gaborszollosy ,

I was recently dealing with the same issue, and in the end I used a workaround. Instead of the conversions I stored the .csv as a .txt file in SharePoint, and when I used 'Get file content' on the .txt file it was formatted as expected.

[gaborszollosy]

Great idea, thanks! So basically if I load a csv from an attachment and create a file with it's content as .txt on SharePoint and use that for the ongoing flow, do you think it would work out?

[tom_riha]

Yes, if the .csv is a SharePoint item attachment then it's exactly the same way I used it.

[takolota]

Does the more direct base64ToString(Attachment Content) expression not work for this?

Also if anyone wants to parse a CSV with commas in the data, feel free to use the template here: https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/CSV-to-Dataset/td-p/1508191

[L3V1A7H0N]

I was just working with this, and saving to a sharepoint library as a .txt file does not remove these characters for me. Get file content pulls back the exact same thing from the .txt file as the .csv. Did something change?

[levicatapult]

Took me all day to figure it out, so hopefully this helps someone. 

Our issue came from the .CSV being created on a MAC computer, and /u0000 was inbetween every character when I converted the contents with the base64toString expression. If you open these files, Microsoft will automatically convert it to a normal windows format. But if you're automating reading that file in Power Automate, you'll need the following.

• Compose1 - stringCSV

base64ToString(outputs('Get_file_content')?['body/$content'])

• Compose2 - uriString

uriComponent(outputs('Compose1')

and then you can replace all the %00 with nothing, and you'll have a normal output from the file. 

[Zulkai]

Thanks so much for this. It's taken me longer than I care to admit, trying to resolve the \u0000 issue.

In the end I didn't need the uriString step. Here is my flow which removes the \u0000 null values from the CSV.

[Staffi]

I spent 3 days banging my head on this one and your solution was the only one to work. Thank you levicatapult.