ArcPro Project Templates-whats the deal?

01-19-2022 09:23 AM
New Contributor II

It's my understanding that a project template is different from a project package in that it doesn't store data but stores pathways and connections and toolboxes etc.

However, my .aptx file is 1.8gigs -- which seems large -- and if I try to include any rasters the tool runs for quite a while and inflates the size dramatically. When I open the project template its pointed to our enterprise data as it should... but it seems dubious that there is THAT MUCH data in the template. 

Is my understanding of a project template flawed? I was expecting its size to be comparable to an MXD in desktop.

11 Replies
MVP Esteemed Contributor

 a template can provide either complete datasets or schema-only data definition

are you making the template for


a specific machine

or to share?

Create a project template—ArcGIS Pro | Documentation

... sort of retired...
0 Kudos
New Contributor II

Sharing it to the server so our editors can make standardized maps. Schema only sounds like what I want but when I was sniffing around I didn't see that option but I'm a bit of a stooge so I may have missed it.

I did not save history items or any of that.

Occasional Contributor

Optionally, check the Include History Items option to include all geoprocessing history items and the data required to reexecute the history item.

If this option is not checked, none of the project's geoprocessing history items are included. If history items are included, ensure that all data required to reexecute the tool still exists and is accessible. If it's not, packaging will not be successful.

you likely need to uncheck include history. that is likely including data you are not intending.

New Contributor II

That box was unchecked.

MVP Esteemed Contributor

You may get a clue of what's going on by renaming your aptx to and vewing it with a zip viewer to see what's inside.

New Contributor III

Other posts / content on this subject have suggested that project templates indeed create a COPY of all the data. Though you mention enterprise, which was one thing that supposedly changed that, along with UNC path. But the standard seems to be that all data is copied (much to the chagrin of many).

I don't really get templates. If you just want to maintain data LINKS and not COPIES, is there functionally any reason not to just have a standard project that gets used as a template (new iterations get 'saved as')?  

New Contributor II

Yeah I don't get it... There are already project packages. And "Save As" creates a bunch of projects pointed to the same default .gdb and stuff like that. Maybe layout files are a better option than project templates? it just seems odd to me that when you create a new project from the template its pointing to the correct data, while a project package is pointing to copy of data. Seems like something inadvertent is happening behind the scenes, why create copies that aren't referenced?

MVP Esteemed Contributor


Can you FYI someone that can follow up on this Project Template vs Project Package issue... it has come up a couple of times 

... sort of retired...
New Contributor II

I'm trying to help figure out a way to get around this for end users in one of our offices, and a solution that will work for other offices that do a similar workflow.

In ArcMap, most of our offices have a series of MXD files, 1 to 10 of them in most cases, that have all of our regularly used layers in them, with official layouts, dynamic text, dynamic legends, all preset.  The user opens the map, zooms to the area of interest, changes a few layers if needed, changes the title, and prints the map. These MXD's are read only, which forces the end users to save it as a new MXD file if they want to keep it for the future and keeps them from changing the clean slate for the next user.

It makes it very easy for most of our users to get a map product out for the record or public use, without having to build something every time.

So, in ArcGIS Pro we want to remake this and I've tried the following...

1. Map File. Doesn't contain layout information, only the map tab, requiring everything else to be done again or a Layout Template to be imported. This is an extra step that is too much for many users.
2. Project Template. This SHOULD be the answer but it completely copies everything into the template file for some unexplainable reason. The result is a corrupt template file that never finishes creating, or a template file hundreds of gigabytes to terrabytes large as it pulls in imagery and everything else.
3. Save As new project file, Read Only.


The best thing I've found so far is #3, but even when you "Save As" to a new folder, as soon as it is opened the first time you get a folder created called "Index" which has two subfolders and is completely empty. This isn't really acceptable due to the location of these files, but I found I can make it a hidden folder and that somewhat works. Being that I've only tested this myself, I have no idea what issues this could bring in the long run with the default GDB's and other project related files from the original project import of the .MXD's, especially when many users start using these.

#2 above should be the correct answer, but considering the data duplication, it's completely unacceptable. Why? We use standardized datasets, some in File Geodatabases, much in SDE, and copying all of hundreds of layers worth of data then makes these orphaned snapshots of the original live datasets. More so, if this wasn't a problem, all of the unneeded data duplication balloons our backups very quickly. Our SDE data is almost completely statewide datasets.

The only way to fix this, that I've seen talked about, is to convert all of the layers to use UNC paths. I've seen no good walkthroughs or tools on how to do this. Multiply this by hundreds of layers in these maps, and you have a nightmare waiting to happen.

There's a big push to get 25% of our users into ArcGIS Pro this year so I want to be able to offer a solution that won't create duplicated data or data that is severed from the origin datasets. I can provide more information as required, screenshots/shares, and the like.