🚀 go-pugleaf

Trevor:
Can you be more specific in what your doing, what PSE7 does, and what you're expecting it to do?
Bob

Bob,

Perhaps Trevor is referring to a feature of PSE 6 and 7 that annoys many of us.

When a search matches a photo in a stack or version set, the entire stack or version set is included in the search results. If the search matches the top photo of the stack/set, it is shown collapsed. Otherwise, it is shown expanded, and photos in the stack/set not matching the search are shown with the red circle-slash.

This behavior makes it impossible to search and select photos meeting a given criteria. For example, someone posted here the following use case: He edited his raw files and saved them in version sets as JPEGs. He wanted to move just the raw files offline. But there�s no way to find and select just the raw files � if you do Find > By Details > File Type Is Camera Raw, the JPEGs in the version sets will also be displayed (with a red circle-slash).

As another example, someone wanted to find and select just those photos with 5 stars, including photos that were in stacks.

In general, what many of us want is the ability to do a search and display exactly all those photos that match, regardless of whether they are in a version set or stack.

Here are some past posts about this issue:

 <http://www.adobeforums.com/webx?128@@.59b6aff4>
 <http://www.adobeforums.com/webx?128@@.59b5c2b7>
John Rolfe Ellis, "Exact match" #1, 30 Apr 2008 10:55 pm </webx?14@@.59b52281/0>
 <http://www.adobeforums.com/webx/.59b50d3e>
 <http://www.adobeforums.com/webx/.59b4b4e2>
 <http://www.adobeforums.com/webx?128@@.3c05cbe1>
 <http://www.elementsvillage.com/forums/showthread.php?p=358500#post358500>

That is exactly my issue. Well put John. I just hope if enough people make enough noise the developers will realise that this is not a logical behaviour of the program. It goes against what people would expect to happen and limits the usability of PSE. I have since unstacked all my pics which is unfortunate.

Aarrrgghhh!!!

This is exactly what I've spent the last 24 hours trying to work out!
I have about 50gb of raw files I need to archive to DVD, but no way of doing it unless I remove all the stacks and version sets I've built up over the last few years!

Oh this is so frustrating! I expect more from a product supposedly aimed at photographers.

I guess I have a choice; I either give up on archiving files or I abandon stacking and versioning. Either way a pretty poor situation. As I run out of disk space I guess it's the stacking and versioning that has to go... in which case I don't want to waste my time (any more of my time!) creating structure in the catalog.

Anyone know of any decent inexpensive alternatives that aren't crippled in this way?

      Anyone know of any decent inexpensive alternatives that aren't crippled
      in this way?




You might look at Lightroom and Microsoft's Expression Media, both of which cost about twice as much as PSE but seem to get better attention to detail and support.

Thanks John.
At present I'm strapped for cash (recently made redundant) so for now I'll have to soldier on with PSE.

There's something very odd happening with stacks... it takes ages to unstack just two photos - around 12 seconds on my Athlon 64 3700+ processor at 2.21 GHz with 2GB ram (not state of the art, but not feeble either), but as I increase the number of stacks I try to undo, I get an exponential rise in the time taken.

After some experimentation I have an approximate relationship - there seems to be a 6 second overhead per batch, then 5.75 seconds for every stack selected, but that total time is increased by about 1/4 of the number of stacks as a percentage - so for 4 stacks that's (6 + 4*5.75)*1.01 (the last term being an increment of 1%) giving 28 seconds, but for 400 stacks it's a lot more than 100 times as long - (6 + 400*5/75)*2 (the last term being 100% increment). It's quicker for me to sit at the computer and manually unstack in batches of up to 10 than it is to ask PSE to do the same number in one go by itself!

It seems to me (I am a software developer working with relational databases) that PSE is being crippled by some very inefficient queries that do not scale - even the 12s for one stack of two photos seems way over the top but for the software to perform worse than a person sitting at the PC and doing it by hand is beyond belief!

I have 7000 stacks to unstack... if I ask PSE to do them for me it looks like it will take around 9 days solid at 100% CPU! Or I could sit there and feed it 10 at a time every minute or so for 12 hours straight! :-{

Interesting � I see something similar. In my largest catalog of 12K items, there are only 167 stacks, but it takes 6-7 seconds to unstack one stack of two photos. I�m guessing that Adobe never thought that such a use case was important to test for speed (it would pretty rare for someone to unstack 7000 stacks!). The underlying database, SQLite, has a good reputation for basic performance, but as you know, it�s how the application uses it.

If you want to be adventurous, you could download SQLite Database Browser (free), make a backup copy of your database file (triple check that!), open it with the browser, and delete all the rows in the tables �media_stack_sort_table� and �media_stack_to_media_table�. I�m pretty sure that would get rid of all the stacks without violating any application invariants in the database.

Thanks again.

I eventually finished the unstacking task. It got quicker as the number of stacks reduced - 400 batches took around an hour and a quarter when I had 7000 batches remaining, with 3000 left a batch of 400 took 26 minutes and the last 400 were through in under 90 seconds (meaning a succession of batches was dramatically quicker than doing it all in one hit!).

I didn't know it was based on SQLite (about which I knew nothing) but I took a look at the database as you suggested. There seem to be a lot of key columns that are, well, not the way I'd have designed them! The evidence of my unstack timings suggests also that each unstacking operation involves looking through the entirety of the stack tables (or something along those lines, I don't have access to the code of course) - possibly joining multiple full-table scans with the selection of the batch of stacks to unstack, which would explain the exponential slow-down for larger batches and for larger total number of stacks in the catalog.

Maybe I should offer to rewrite the data access layer for Adobe? ;-} I need a job right now.

Glad you got it finished. (I'd guess there's probably an index missing on one of the stack-related tables.)