GP SoftwareTwitter
Opus FAQsManualCommandsObjects

Bug report: duplicates aren’t really sorted


#25

Thank you, leo, and looking forward for the result.


#26

Leo, any news about this problem?


#27

No. (We have a huge to-do list. We'll update the thread and/or put something in the release notes when there's anything to announce.)


#28

Just wanted to add my 2 cents. It seemed to me that in prior versions of DO when I would search for duplicates across folders by checksums that the results returned were ordered by the way the directories were included in the initial list to search across. If you listed three directories (XXX, AAA, RRR) you would get your duplicates of a single file with all dupes in XXX, then AAA then RRR regardless of the file name.

file r in XXX
file t in XXX set for delete
file t in AAA set for delete
file p in RRR set for delete
file t in RRR set for delete

All the above would be dupes of one another but it was never random as to how they displayed.

That seems to have changed!

Today when I did a dupe finder, sometimes I would get the RRR first even though there were dupes in AAA and XXX which used to be listed first or as I said by the order the directories to be searched were added. I know this use to work because I typically would compare a new directory against several older directories and as long as I setup the new directory (I can't remember anymore if it was first or last) it would always auto delete the newest files and I wouldn't have to time going through them.


#29

Almost forgot. The whole reason I was searching on this subject was because I noticed the change in the order and came to see what exactly was being used as a tie breaker to determine the order.


#30

[quote="cityguy"]Just wanted to add my 2 cents. It seemed to me that in prior versions of DO when I would search for duplicates across folders by checksums that the results returned were ordered by the way the directories were included in the initial list to search across. If you listed three directories (XXX, AAA, RRR) you would get your duplicates of a single file with all dupes in XXX, then AAA then RRR regardless of the file name.

file r in XXX
file t in XXX set for delete
file t in AAA set for delete
file p in RRR set for delete
file t in RRR set for delete[/quote]

This thread is about how the duplicate groups are sorted.

How the files within the groups are sorted has never changed:

[ul][li]They are sorted according to how the file display was sorted before you ran the duplicate finder.[/li]
[li]They can also be re-sorted by changing the file display's sort order and clicking the Select button near the bottom of the Duplicate Finder panel (assuming you are in Delete mode).[/li][/ul]

You may need to add the Location column and sort by it.


#31

I know it's about how the data is sorted within the groups. That's exactly what I noticed had changed. Before it appeared ordered by the way I entered the directories and now it appears random.

I haven't changed anything in my setup. I'm not sure why the results have changed but like I said I never had to do anything special before other than make sure my directories were specified in the correct order prior to starting the process. I've never clicked on the "select" button, didn't even know what it's purpose was. I'll give it a try next time to see what it does and see if it solves my problem.


#32

The thread is not about how the data is sorted within the groups.

It's about how the groups themselves are sorted, not how what's inside each group is sorted.


#33

This has gotten very complicated for me but I would like to repeat a couple of points that Leo made that I stumbled upon myself and which help me with my dupe finding

[quote="leo"]
How the files within the groups are sorted has never changed:

[ul][li]They are sorted according to how the file display was sorted before you ran the duplicate finder.[/li]
[li]They can also be re-sorted by changing the file display's sort order and clicking the Select button near the bottom of the Duplicate Finder panel (assuming you are in Delete mode).[/li][/ul][/quote]

The first point is the most important. Changing the file display before finding Dupes greatly effects your results. Another trick that might come in handy, when finding Dupes below ONE folder, try using Flat View. It can help you understand the total pool of data that you are working with and which way of pre-sorting works the best.

I think Leo's second point was lost on some. Once you have your list of duplicates and you don't like the choice that DOpus has made for deleting (checkboxs), sort the dupes a different way and then click the Select button. DOpus will then re-select the checkboxes, always leaving the top-most one as the "keeper".


#34

Well, the latest developments (10.1.0.3 and 10.2.0.1) make things more difficult:

  • 10.1.0.3 reverted the order of date and size fields. I don't know whose idea this was, but it was a bad idea in general in my opinion. Usually, when I sort my files, I want to have them in chronological order or from the smallest to the biggest. But with duplicates, it's worse, because if you sort them by date, DOpus will automatically select to keep the most recent find instead of the oldest. Every time I have to click twice on the date or size fields. In the past there was an advanced option that allowed you to revert the sort order. It's a shame this option wasn't even proposed as an alternative for the 10.1.0.3 change.

  • 10.2.0.1 "Added an option to the Duplicate File Finder to number the duplicate groups rather than using the search parameters to describe the groups" Maybe this is an incomplete try to allow users to sort duplicate groups as they want (This was the reason I created this thread.) But, we don't really know how the digits this option created are useful. Moreover, after I used this option once, now it's gone and it doesn't appear anymore on new duplicate searches!


#35

You can change the default sort order from Preferences / Display / Fields.

What do you mean the option is "gone" ?


#36

What a relief! I didn't see this information in the change log. Thank you

I turned on "Folders / Folder Display / Add 'Group' column automatically when file display is grouped" and "number duplicate groups" in the duplicate search panel, but the group column doesn't appear anymore in the duplicate tab when I perform the duplicate search.


#37

That's because it hasn't changed :slight_smile: You've always been able to reverse the default sort order of each field that way, all that changed was the default sort order of a few fields.

If the Group column has been turned off somehow you can turn it back on in the same way you can turn any column on (right-click on the column headers, or use the Folder Options dialog).

The option to add the column automatically only takes effect when the display as actually set into group mode; if you have, for example, a folder format saved for a collection with grouping turned on but the Group column turned off, the option won't have any effect.


#38

I was able to have the group column back for a while, but it looks like it doesn't stay there between duplicate searches. You have to set it manually almost every time, regardless of the "number duplicate groups" checkbox.

Anyway, it's still not clear to me what this group digits represent (see the screen capture). It's obviously neither the size nor the MD5.



#39

You probably just need to save the folder format for the Duplicates file collection (see the Folder Formats FAQ for details).


#40

Okay, I will try that.

What about to the digits appearing in the group column in the screenshot above? How are they useful?


#41

Yes that will be improved in the next version (remember that 10.2.0.1 is a beta release).


#42

Well, things are getting better in 10.2.0.2. Thank you so far!

But I still encounter some bugs in this version:

  • sometimes after you've just run a duplicate search, columns are not in the right place (see the screen shot below)
  • files from the root of the library are located as such whereas files from its subfolders have the real path, making it more complicated to sort by the location field
  • if you click on column headers and / or group by columns, groups of duplicates could mix between them, so you don't recognize any more which file is the duplicate of which one (see the second screen shot)




#43

In the first screenshot, I agree about the location field showing the library root - that should be fixed.
The column headers being wrong is very weird. Do you ever see that effect outside of Duplicates? If you remove a column and re-add it does it fix itself up?

In the second screenshot, you're grouping the results by Name, not by Duplicates.


#44

I’ve only seen this a few times in the duplicates since the new group column was created. But it might not be related to the duplicates, because I only feel the need to group files in the duplicates.
I didn’t identify how this column shift appeared, but I think I encountered it when I re-run the same duplicate search while being in the duplicate tab already.

As far as I remember, columns start fixing themselves if I start clicking on them. Next time I see this I will try removing and adding columns, as you suggested.

OK, I think I got a bit confused about this new column and forgot about the special group by duplicates. But I wonder if duplicates shouldn’t always stay together in some way.