Importing CSV file - encoding

martinsirko

Hi.

I am trying to migrate from my previous password management system to 1Password. I tried to import CSV file, but non-ASCII characters are not imported correctly. Is there some way how to solve this issue.

I have also try to create a record in 1Password with non-ascii characters, export it to CSV (the results were correct), and than reimport it back to 1Password, but unfortunatelly the non-ascii characters weren't imported correctly. So I assume it is not possible to import non-ascii characters correctly, but hopefully I am wrong.

Is there any known solution to this problem?

MrC

Hi @martinsirko‌,

Most likely your file was not saved as UTF-8, the encoding that is required for 1Password to correctly decode the data.

Can you mentioned which password manager you are using? There may already be a good solution for you.

martinsirko

Hi.

The file is encoded in UTF-8. The problem is that also when you export passwords from 1password and import them back, the encoding will get broken. I tried that on Windows.

I have imported my passwords and edited them manually, but maybe this is something you could try to fix in the future for new customers.

MrC

Hi @martinsirko‌,

Maybe someone from the AgileBits Windows team will provide feedback - I haven't used the 1Password CSV importer or exporter in a while now, so don't know what state it is in. I know a few bugs were fixed earlier last year, but once I migrated my converter utilities to export to 1PIF, I know longer personally concerned myself with the CSV importing.

svondutch

The 1Password CSV importer supports ANSI and UTF-8. I'm afraid it doesn't support any other character encoding at this moment.

The problem is that also when you export passwords from 1password and import them back, the encoding will get broken. I tried that on Windows.

@martinsirko‌ Do you happen to have consistent, reproducible steps? I would to reproduce and fix this problem.

miyazawa

I have same trouble.

This is reproducible.

1. Add an new item, such as "Login"
2. Use 2-bytes character on "Title", "Username", "Notes"
3. Export this item to csv
4. Inport this item, and most "Title", "Username", "Notes" become different.

And I find not perfect solution.
Befoe importing csv file open csv file with notepad, you can see correct character, then overwrite.
The import.It's much better.
I compare 2files, but there are no difference.
I wonder why this causes better results.

And this also happens when exporting and importing 1pif file format.

2-bytes character sample is below.
"Title" ==> "日本語"
"Username" ==> "氏名"
"Notes" ==> "メモ"

I hope this problem can be solved.

MrC

The encoding does get botched. Here's an example and a UTF-8 test file:

Test file: https://dl.dropboxusercontent.com/u/87189402/test_utf8_multi.zip

svondutch

The encoding does get botched.

@MrC @miyazawa‌ @martinsirko‌ Fixed in 4.1.0.BETA-534

miyazawa

Wow!!
It's great work!!

I can import from keppas db that has many 2bytes charactes by csv.
Thanks a lot!!

MrC

FYI @miyazawa‌,

You can convert from a Keepass 2 or KeepassX export. See this thread.

MrC

@svondutch,

There seems to be another encoding problem w/4.1.0.538. The left side data was entered into Notes and the right side results after import:

Create an entry, and add this to the Notes:

These are some multiline notes:

with special " chars

à ß ÿ ą Ŵ œ

END

export the record as a PIF, and import it.

svondutch

@MrC Thank you for bringing this problem to my attention. I have fixed this bug in version 4.1.0.BETA-539

MrC

Confirmed.