Author Topic: Strange image numbering in search results  (Read 372 times)

PaulS

  • Full Member
  • **
  • Posts: 142
Strange image numbering in search results
« on: December 27, 2020, 01:42:12 AM »
Search Result windows that are created by searching for Duplicates (or Copies (binary identical)) have strange file numbering when "Include originals in results" is selected.

For the results below, I searched with 1100 originals and found 1636 duplicates:

"Include originals in results" Disabled:
  - total number of files in header is 1636 (duplicates only)
  - files in the window are numbered from 1 to 2736 (originals + duplicates)

"Include originals in results" Enabled:
  - total number of files in header is 2736 (originals + duplicates)
  - files are numbered from 2 to 3836 (2 * originals + duplicates) even though there are only 2736 total files

This is not a major issue but it is a bit unexpected and confusing.

Mario

  • IMatch Developer
  • Administrator
  • *****
  • Posts: 28532
Re: Strange image numbering in search results
« Reply #1 on: December 27, 2020, 12:41:42 PM »
What do you mean by "files are numbered"? Which number do you refer to?

PaulS

  • Full Member
  • **
  • Posts: 142
Re: Strange image numbering in search results
« Reply #2 on: December 27, 2020, 01:16:39 PM »
The number that shows on the thumbnail.

Mario

  • IMatch Developer
  • Administrator
  • *****
  • Posts: 28532
Re: Strange image numbering in search results
« Reply #3 on: December 27, 2020, 01:38:43 PM »
I understand you have configured your File Window to show the file index?



In that case: If the originals are shown separately (in the header), the first original gets index 1. The first match for this original gets 2 and so on.
If the original is displayed alongside the matches, it gets index 2 (the 1 is not used in this case). The same is true for all other originals. Even if the originals are shown "inline", the formerly allocated index remain in use - you get some gaps in the index sequence.

When I get produce a result with 2 originals and 50 matches each, the index runs from 1 to 102 when the originals are shown in default mode. And from 2 to 104 when the originals are inline.

I would not consider this as something that has to be changed...? I will think about it.
« Last Edit: December 27, 2020, 01:53:58 PM by Mario »

PaulS

  • Full Member
  • **
  • Posts: 142
Re: Strange image numbering in search results
« Reply #4 on: December 27, 2020, 05:47:04 PM »
Yes, you are exactly right.

The original gets an index number assigned whether or not the header is shown, and gets a second index number assigned when it is included in the results.

In your case it isn't so noticeable since you had only 2 originals.  In my case where I selected >1000 originals and found many duplicates it is noticeable.

I agree that it is not a major issue, especially since I now understand and expect it.  But it was really confusing initially.

[FYI so you understand my unique use case for finding large numbers of duplicates:  My wife likes to make calendars and photobooks for family and friends, each of which can include 10's to 100's of photos.  Although this is an ideal task to perform in IMatch, she is a bit intimidated by computers and so uses a process which she considers simpler and easier:  1) create new folder for each calendar/ photobook, 2) copy the photos she likes to the folder, 3) upload them to a photobook service.

Since these groupings are also interesting to me, I like to keep track of them for future reference under an "Albums" category (eventually I will investigate using Events instead).  The Duplicates feature makes this fairly easy because I can temporarily add one of my wife's new folders to my database, perform a Duplicate search, add the search results to a child category to Albums, and then delete her folder.]

PaulS

  • Full Member
  • **
  • Posts: 142
Re: Strange image numbering in search results
« Reply #5 on: February 04, 2021, 02:53:47 AM »
Quote from: Mario, Version: 2020.14.2 Release Notes
When displaying originals alongside the matches (instead of in the separate group header), the index attribute which you can display in the thumbnail panel runs from 2 (for the original) to n. IMatch now uses index 1 for the first original in this case.
 Change | #01322

Thanks for the fix for this in 2020.14.2.  Works great!