Migrating from Dashlane back to 1Password - is there a clean way of merging the two databases?

ontheroad

I migrated very unwillingly in April 2014 from 1Password Windows to Dashlane due to technical issues that I encountered in 1 Password at the time and which the support team was unable to resolve. I recently discovered the root cause and as a result am now planning to migrate back. Of course during that time, I updated some passwords, added some websites (I currently have approximately 550 items in Dashlane) and updated notes. Complicating the issue is the fact that when I migrated to Dashlane it did not pull everything over from 1 Password (more than 700 items in 1Password) and I never completely migrated over the missing records. So rather than just do an export from Dashlane and a clean import into a new 1Password database, I need to merge two databases.

My first thought was to create a .csv export from Dashlane and use a script posted in the forums here to clean up the resulting export file and then import it into 1Password. However, my expectation is that doing so will give me approximately 1250 records that I will have to wade through to manually compare and delete the duplicates.

Consequently, my questions are:

1. Is there a way to filter out and delete those records that are identical in every respect?

2. Is there a way to merge records where there may be differences in only 1 or 2 fields, eg. passwords and/or notes

3. if yes to question 2, can I select which field should be used?

DBrown

Your "CSV-export, CSV-to-1PIF-conversion, 1PIF-import" method seems like the best path. You may, indeed, end up with duplicates in your 1Password vault.

Automatic detection of duplicate items is already on the list of requested features, but I know of no resources dedicated to development of that feature at this time.

You could sort the list area by Name or URL (probably both, in a couple of passes) to scan visually for repeated items, so you can decide which ones to keep and which to delete.

MrC

Since other password managers allow same-named fields, de-duping in general is non-trivial and ambiguous.

plasma

Will the convert_to_1p4.pl script be updated with a converter for Dashlane? Or is one of the included converters already compatible with Dashlane?

I tried using the lastpass converter on a Dashlane export csv and got 40 "Untitled" notes that I couldn't access* and no logins.

*this might be due to having an expired trial of 1Password 5. I don't really want to buy the software until I can be reasonably sure I'll be able to import my 461 logins and 40 notes, though.

MrC

Hi @plasma,

See my reply here and follow the link I mention. I explain the issues with Dashlane's export. My exploration was back in August 2014, so maybe things have improved with Dashlane's CSV export since then, but their release notes only mentions one release which had unnamed bug fixes. I can take a look at their export.

Which OS are you running Dashlane on ?

plasma

@MrC ,
First of all, I should say thank you for all your hard work on this utility.

Secondly, oh goodness. So the only solution is manual correction of 500+ items in Excel/Numbers.
I really expected better from Dashlane. Their product is otherwise so user-friendly. Sigh.

Thanks again for your help!

MrC

Let me take a look at how good the CSV export is with the most recent version of Dashlane. Which version of Dashlane are you running, and on which OS?

MrC

@plasma - I just took a look at CSV export for the most recent version of Dashlane for Windows. Unfortunately, it is still very broken, and has all the issues I mentioned in the link provided in the post above. In fact, I'll go so far as to call their CSV export Complete Garbage.

The Excel export they provide is better; however, it too suffers from the lack of per-row descriptions of the data. It is almost impossible to reliably determine what type of data is in a row. I suppose some heuristics could be employed to guess at the category, based upon the some characteristics of the data in each field for the record, but it won't be terribly reliable.

What are your predominant categories (types) of records in Dashlane?

plasma

@MrC - Excel export? Is that a Windows option? The only options I see under OS X (Yosemite 10.10.3) are "Dashlane secure archive..." and "Unsecured archive (readable) in CSV format..."

I have 461 Passwords (organized in 13 categories), 40 Secure Notes (in 9 categories), 13 Payments (bank accounts, credit/debit cards, PayPal accts), and 56 Receipts (in 3 categories).
I don't mind losing the receipts and copying payment info manually. The logins and secure notes would take minimum half a day to copy manually though, so I'm really hoping there's a more efficient way of moving those over.

MrC

@plasma,

Yes, that's a Dashlane for Windows export option (which is why I asked which platform you were on, since we're posting in the Windows forum). At least this data export format does not suffer from what I describe next...

The CSV export does not quote CSV-metacharacters. So a password entry, with a notes section of:

Multiline, special chars, ends with a quote "
â Ĉ Æ Õ ɮ
double quotes "stuff" and ""
END

gets exported as:

"Joe Blow","example.com","me@example.com","supersecret","Multiline, special chars, ends with a quote "
â Ĉ Æ Õ ɮ
double quotes "stuff" and ""
END"

The double quotes in the notes field should be doubled up, but they are not. So it makes the end of the cell ambiguous, especially with the embedded commas. Here, for example, are two different ways two spreadsheet programs interpreted the data (Numbers 3.5.2 and LibreOffice 4.3.4.1):

I'm not exactly sure how the CSV parser used by the converter will handle this botched CSV.

Edit: I tested the CSV parser - it can't do much better. Even when I supply it with options to try to handle broken quoting, it is easy to trip it up with certain data. And since there is nothing in the row of data that indicates the type of data (which might imply how many fields to expect), and because the data can contain newlines, there's no way to know where the broken quoted data should end.

MrC

Update: DashLane 4.0 is no better.

See https://discussions.agilebits.com/discussion/comment/273495/#Comment_273495