Versioning master folder etc?

[Ferdinand]

Peter

I don't pretend to understand all of this either.  I understand the problem, and it's a question of finding what works.

You may well be able to specify a higher level folder, but you will need to be sure that you've configured the version relation to search down enough levels.

You may also find what was discussed in the other thread - that new versions added in later may not get the correct relation - master or version, until you refresh version relations for the whole DB.

In terms of your problem with the folders being at the top level, I tried this and it worked for me.  I suggest you check all your settings and run F4,R.

Ditto for the original files on an external (USB) drive.  This worked as well.  Not sure what to suggest other than rechecking

[cytochrome]

With an image tree like

Raw
- JPG
- Web (JPGs also)

I can update (selected) metadata tags from JPG to RAW. The rule defines JPG as master, I start in the JPG folder with Where to search = Master files and Version= 1 level up.

When it is done one has to reset immediately everything to normal (normal is personal of course, for me it is from RAW = master to 2 levels down = Versions and not "all XMP" in what to copy to versions).

I must say that I am puzzled by all this discussion, the rule definition system in IM as it is now is not simple (but can it be?) but very flexible.

Francis

[DigPeter]

Peter

I don't pretend to understand all of this either.  I understand the problem, and it's a question of finding what works.

You may well be able to specify a higher level folder, but you will need to be sure that you've configured the version relation to search down enough levels.

You may also find what was discussed in the other thread - that new versions added in later may not get the correct relation - master or version, until you refresh version relations for the whole DB.

In terms of your problem with the folders being at the top level, I tried this and it worked for me.  I suggest you check all your settings and run F4,R.

Ditto for the original files on an external (USB) drive.  This worked as well.  Not sure what to suggest other than rechecking

Thanks Ferdinand.  I will continue to hack around   I have suggested in a feedback to the Help topic that an example of a pattern for a split internal/external HDD database would be helpful.

[DigPeter]

I must say that I am puzzled by all this discussion, the rule definition system in IM as it is now is not simple (but can it be?) but very flexible.

Francis, it is that lack of simplicity which is causing the discussion.  I am not suggesting it should/could be simpler, but for a non-teccy, like me, it is difficult.  Also, I keep my originals/masters on an external HDD and versions on the internal drive, which seems to complicate matters.

[Ferdinand]

With an image tree like

Raw
- JPG
- Web (JPGs also)
........
I must say that I am puzzled by all this discussion, the rule definition system in IM as it is now is not simple (but can it be?) but very flexible.

There are two things you can do to avoid the problem that Peter and others are having.  One is to have all versions in sub-folders, as you have.  Another is to ensure that no two files are identically named.  I made both changes some years ago.  The change to my folder structure was a large and somewhat risky change, but I'm glad that I did it.  I could see at an early stage that it was going to make version relations easier and faster, but I haven't realised by how much.

[Ferdinand]

I will continue to hack around   I have suggested in a feedback to the Help topic that an example of a pattern for a split internal/external HDD database would be helpful.

If you can post another screen grab that shows this file structure and your settings I can make another attempt to replicate it.  I though I had, but perhaps not.

[DigPeter]

If you can post another screen grab that shows this file structure and your settings I can make another attempt to replicate it.  I though I had, but perhaps not.

Ferdinand - Here it is - thanks.

[attachment deleted by admin]

[Ferdinand]

And a screen shot of the version relations configuration too please.

[DigPeter]

And a screen shot of the version relations configuration too please.

Sorry - now attached.  I tried the pattern you advised.  This worked for the database that was entirley on the internal drive, but not for masters on an external HDD.  I played with other configurations of {p} & {d} without success.

[attachment deleted by admin]

[cytochrome]

I think that there is not enough information in your Master and Link expressions, they are redundant as far as JPG is concerned.

One needs a naming convention if one wants to define at the same time JPG as master and version. I use a simple scheme

- name = RAW
- name_ASP, or name_BIB or name_DxO etc with possibly numbers like name_ASP-1 are all JPG. This, plus the suffix JPG, NEF, RW2, MRW is enough to define un-ambiguous relation rules.

Also, a folder structure which contains implicitly the distinction between RAW and JPG/TIFF is a great help to set the directions of search. I put RAW at the upper level and derivatives below. Also when I have only JPGs (I have a Finepix F30 that does no RAW) I put the master JPGs in a level and all JPG versions derived in lower levels.

Its two months since I had no versioning problem.

Francis

[sinus]

With an image tree like

Raw
- JPG
- Web (JPGs also)
........
I must say that I am puzzled by all this discussion, the rule definition system in IM as it is now is not simple (but can it be?) but very flexible.

There are two things you can do to avoid the problem that Peter and others are having.  One is to have all versions in sub-folders, as you have.  Another is to ensure that no two files are identically named.  I made both changes some years ago.  The change to my folder structure was a large and somewhat risky change, but I'm glad that I did it.  I could see at an early stage that it was going to make version relations easier and faster, but I haven't realised by how much.

I want stress this statement of Ferdinand, although I have masters and versions in the same folder, but no two files have identically names.

[Ferdinand]

I assume the selected folder in the relation definition is correct Peter, and that IMatch media and folders isn't showing the full tree.

Yesterday it seemed to work easily, but not so today.  After I relocated the originals to a thumb drive, it wouldn't pick up the originals as masters.  If I selected all the "original" files on the removable drive and refreshed relations then it would correctly detect them as masters.  Another way was to turn off that relation definition, F4,R, turn it back on again, and F4,R again.

The real challenge is in adding a new version.  You may get the new versions as both master and version.  You will need to repeat one of the above steps to get the version relations correct.

I think you're going to have to try things like this and and find an approach that works for you.

[DigPeter]

I assume the selected folder in the relation definition is correct Peter, and that IMatch media and folders isn't showing the full tree.

Yesterday it seemed to work easily, but not so today.  After I relocated the originals to a thumb drive, it wouldn't pick up the originals as masters.  If I selected all the "original" files on the removable drive and refreshed relations then it would correctly detect them as masters.  Another way was to turn off that relation definition, F4,R, turn it back on again, and F4,R again.

The real challenge is in adding a new version.  You may get the new versions as both master and version.  You will need to repeat one of the above steps to get the version relations correct.

I think you're going to have to try things like this and and find an approach that works for you.

Thank you very much for looking at this.  It too works for me when masters and versions are on the same internal drive, but not, as you found when the masters are on an external drive.  All settings are the same for both cases, except for where the masters are held.  As the versions are held on the internal drive, the "processed" folder containing these is the same in both cases.  It is the folder that is selected when refreshing with F4,R.

As nobody else has suggested a solution, I am hoping that Mario will in good time, show the way.

[DigPeter]

I think that there is not enough information in your Master and Link expressions, they are redundant as far as JPG is concerned.

One needs a naming convention if one wants to define at the same time JPG as master and version. I use a simple scheme

- name = RAW
- name_ASP, or name_BIB or name_DxO etc with possibly numbers like name_ASP-1 are all JPG. This, plus the suffix JPG, NEF, RW2, MRW is enough to define un-ambiguous relation rules.

Also, a folder structure which contains implicitly the distinction between RAW and JPG/TIFF is a great help to set the directions of search. I put RAW at the upper level and derivatives below. Also when I have only JPGs (I have a Finepix F30 that does no RAW) I put the master JPGs in a level and all JPG versions derived in lower levels.

Its two months since I had no versioning problem.

Francis

Francis,  thanks.  I understand that, but my problem is that I wish to keep my originals (RAW & jpg) on an external drive and the versions on the internal drive.  Versioning works for me when all files are on the internal drive, but not when the masters are on an external one.  Having files with the same name does not affect this issue, though I agree that it is desirable that names should not be duplicated.

[BenAW]

As the versions are held on the internal drive, the "processed" folder containing these is the same in both cases.  It is the folder that is selected when refreshing with F4,R.

What do you precisely mean with " It is the folder that is selected when refreshing with F4,R."?

[DigPeter]

As the versions are held on the internal drive, the "processed" folder containing these is the same in both cases.  It is the folder that is selected when refreshing with F4,R.

What do you precisely mean with " It is the folder that is selected when refreshing with F4,R."?

I select the folder that contains the versions when I do F4,R.  In the structure which I posted in response to Ferdinand (reply #56), this the Folder entitled "Processed".  I realise that this is not the way you do it, but it works for me when the masters are on the internal drive with the versions, but not when masters are separated onto an external drive.

[BenAW]

To be perfectly clear: you select ONLY versions, NOT masters when you do F4,R?

[DigPeter]

To be perfectly clear: you select ONLY versions, NOT masters when you do F4,R?

All masters are in the "Originals" folder, so, yes, only versions are in "Processed"

[BenAW]

How do you expect IM to process masters if you don't select them?

[DigPeter]

How do you expect IM to process masters if you don't select them?

OK then back to your question: "What setting is needed to have only files in the 'Originals' folder shown as master?"
Only select files in the Originals folder when you do F4,R.
With the correct settings IM should only look in the Processed collection for matches.

As I said, it works when everything is in the same internal drive.  I was following what I understand you to say your earlier post .

[Ferdinand]

Peter - when I said "If I selected all the 'original' files on the removable drive and refreshed relations then it would correctly detect them as masters", I wasn't clear on how to do this.  How I did it was in the media and folders toolbar, turn on "Show Files with Hierarchy" and in the little drop-down arrow next to this button, choose "Show All Levels".  This will enable to you see all the originals at once when you select G:\test7\originals, and so to select them all, and then to F4,R.  I think you will find that if you do this it will work, but you may need to do it after each ingest.

[BenAW]

As I said, it works when everything is in the same internal drive.  I was following what I understand you to say your earlier post .

To establish relations you have to select a set of images and do F4,R (if automatic versioning isn't kicking in).
Simplest way is to select @All and filter OUT the folder containing versions.
Do CTRL-A to select all remaining images, or only a selection to speed things up.
Now IM ONLY gets images that are NOT a version, and IM can try to find versions for each file.

Throwing versions only at IM is never going to work.

[Ferdinand]

Throwing versions only at IM is never going to work.

In my tests trying to simulate Peter's set-up it did work sometimes but not always, and definitely not with the split drive arrangement that he's using.

[DigPeter]

Oh dear, what a tangled web - my head is spinnning 
I cannot spend more time on this now, so very many thanks Ferdinand & Ben.  I do appreciate your efforts to help this computer illiterate.  Not being a professional photographer, I probably do not have the same imperatives. My objective is to have the same User Metadata (e.g. title, description, keywords) in masters and originals.  I am now working on a workflow to achieve this without versioning.  I will return to the problem in the future.

[DigPeter]

Lol or Col
Having said I would leave this alone, I have inadvertently created versions in the same database that we have been discussing, without F4,R.   This is what I did (see attachment):

Select @all 
Select a version and copy "My metadata " - this image is the internal  "C:\...\Processed folder
Select the master of the previous image and paste "My metadata" - this image is in the external G:\...\Originals folder
After IM has done its updating stuff, lo and behold, versions have been created.

The file relations are those attached to reply#58
( https://www.photools.com/community/index.php?topic=1414.msg8768#msg8768 )
So I then selected all images (that is both masters and versions) in @all, did F4,R and all versioning was done correctly. 

Fluke? Design? 

The same relationship was used as shown in reply#58 ( https://www.photools.com/community/index.php?topic=1414.msg8768#msg8768 )

[attachment deleted by admin]

[BenAW]

This is the result of a overcomplicated versioning setup. Pity.
Just tried in a test setup with masters on C: and versions on D: disk, works of course as expected.

[DigPeter]

This is the result of a overcomplicated versioning setup. Pity.

What do you mean?
It seems to me to be very simple - masters all in one place and versions in another.

[BenAW]

This is the result of a overcomplicated versioning setup. Pity.

What do you mean?
It seems to me to be very simple - masters all in one place and versions in another.

I mean the way versioning is implemented in IM, it is unnecessarily complicated.
I have been using and promoting the split setup for masters and versions the last couple of years, because it works.

[DigPeter]

I mean the way versioning is implemented in IM, it is unnecessarily complicated.
I have been using and promoting the split setup for masters and versions the last couple of years, because it works.

Ah - I thought you meant my set up - apologies.  

[Ferdinand]

Fluke? Design? 

Perhaps both, perhaps neither.  I suspect that you may not get this result consistently.  In my testing there were a couple of ways you could get the correct result consistently with your folder structure and version relations configuration, discussed above.

I think that if you were able specify the precise folder that each version was in, like in my "{p3}\Processed\{d1}\{d0}" suggestion, then that would work without having to be careful what you were selecting when you did F4,R.

It is true that there is some complexity in version relations configuration.  IMHO this is because it tries to accommodate a variety of workflows.  This is always the IMatch principle.  Sadly despite this there is one that it doesn't support, or at least not easily.  I can't agree that simply because it doesn't support this one that therefore it is too complex.  As enthusiastic as Ben is about his folder structure and workflow, it isn't going to suit everyone.

I have come to Herman's view that having an *option* to exclude files from being masters if they're in one of the folders that is specified as a place for versions would be a good idea.  Yes, this would add more complexity, and for that reason Mario is unlikely to agree, but it would be consistent with the principle of supporting multiple workflows.

[DigPeter]

@Ferdinand - thanks.  I will return to this in due course.

@All - Meanwhile Season's Greetings to all.