Suggested Survey123 Schema Length

Question

Hey there,

Out of curiosity what would be considered as too big/much? I have a survey I created with connect and it has 1837 lines of calculations and fields that also pull data from 6 other excel sheets. Now I know that this works but it takes a longggg time to load up and submit surveys. I normally use power automate for my reports but I get an error that my schema is too big and exceeds 1024 although there's really "only" 357 fields in the actual feature, there's 1837 lines to the survey including the hidden ones. so I'm assuming power automate is talking about that. But anyways back to the question I'm just curious what seems to be the line of "wow your really pushing it" vs "the software appreciates your minimalism"

abureaux · Answer

Beyond calculates, anything that adds a related table also drags down operations. I have a survey with ~50 repeats and it can be brutal to work with. I think it has around 1500 lines as well, but the lines don't concern me, whereas the related tables do.

dougbrowning · Answer

Depends on number of calcs but for me I found that around 4-500 and the tablet starts to struggle with memory.  Older versions of 123 the app would just shut down on us.

What I do it split it into several forms and then relate them all to a main point.  One reason is we can have multiple users with multiple tablets on the same site.  Not sure if that would work for you.  My workflow for that is here  https://community.esri.com/thread/221263-mapping-with-survey123-within-a-polygon-or-admin-unit

But first some ideas to help the load.

Are you setting the bind::esri:fieldType to null for all intermediate fields that do not need to be stored in the data?  I also wonder if you can create a view with less fields and then use that for the power automate.

Are you breaking the form into groups and adding relevant to these?  That way the entire form does not load all at once.  Esp on the pulldata these do not need to load until a value is picked.  I have found that pulldata tends to recalc a lot also.

Are you using once() or calculationMode=manual or whenempty where you can?

Are your lists external using select from file or the search appearance?

Do you have newer tablets with lots of free space?  Also are they higher end?  We have tested all the iPads and a iPad Pro loads significantly faster than a regular (and an Air is in the middle).  Only time we had speed complaints we found out teams were buying older gen model refurbs that were sometimes 5+ years old (which with Apple's crazy naming is easy to do).

If you can post a version of the form that would help.

Hope that helps