topic Re: Counting "unique" selections across repeats in ArcGIS Survey123 Questions

Counting "unique" selections across repeats

KarenTuerk2 — Wed, 10 Apr 2019 23:33:02 GMT

OK, here's the deal. I am trying to create a field that will count the total number of plant species observed in a plant survey. The survey has 100 subplots (repeating form 1-100). Each subplot repeat and has a nested repeat that asks observers to choose all the plant species found in that subplot. I can use count(${plant_sp}) to get the total number of plant species selected throughout the survey, but I want to count only the number of unique species selected in the survey. This is to find the species richness of the site.

For example, if there is a daisy, coneflower, and hawkweed in every subplot, I need the count to be 3 and not 300. There are literally hundreds of possible choices for each subplot.

Is there a way to pull the number of unique codes from a concatenated list of all the species codes #from a
join(' ,' ${plant_sp}) perhaps?

Thoughts?

Thanks!

Karen

Re: Counting "unique" selections across repeats

MaazaMekuria — Thu, 11 Apr 2019 01:15:22 GMT

Not clear if you want one for a subplot or one per overall survey? You may have to resort to Python or other programming tool after the data is stored if you would like to get one per sub-plot. If per survey, you can make a variable for each flower type and compute a variable if it was selected at least once its count would be greater than zero (set the default value to 0). See below a use case for summing repeats from https://docs.opendatakit.org/form-repeats/#referencing-answers-in-repeats . Also you can use the selected function for each item to find if present in the list of selected item.

Some functions such as indexed-repeats are unfortunately not implemented in Survey123.

Re: Counting "unique" selections across repeats

KarenTuerk2 — Thu, 11 Apr 2019 14:31:43 GMT

Thanks Maaza,

I need to find unique selections across the survey. The number of selections (plants) will always match the number of species within the subplot. But the same species map be chosen more than once across the survey.

Unfortunately, setting a variable for each species is not practical as there are thousands of plant species choices.

Karen

Re: Counting "unique" selections across repeats

JamesTedrick — Tue, 23 Apr 2019 21:52:29 GMT

Hi Karen,

This is something Survey123 may be able to support in the future - in the Early Adopter Community there is a feature 'custom JavaScript functions' that has been enhanced to be able to process repeat information. The specific function would be dependent on the form's layout.

Re: Counting "unique" selections across repeats

DougBrowning — Tue, 23 Apr 2019 22:32:27 GMT

You are on the right track. Here is how I did it. Not perfect.

Create a large txt (or note) field outside the repeat. i did text so users can copy and paste it.

Add a calc joining the fields from the repeat join(", ", ${SpeciesList})

Set a constraint on this join field (not the field) - the trick here is to remove the last entry since it is the one you are trying to add currently.

So like this not(contains(substr(${AllPlantsCheck}, 0, string-length(${AllPlantsCheck}) - 3), ${SpeciesList}))

--If you do not do this it immediately adds to the list then complains it is already there.

--(note my plant codes are 4 chars. so if i remove the last 3 chars it will not find what I just added)

It can be funky if you go backwards and then change something but it still seems to work ok.

I even built on exceptions to the rule. If it contains a XXXXX generic code then the repeat is allowed.

Like contains(${SpeciesList}, "XXXX") or not(contains(substr(${AllPlantsCheck}, 0, string-length(${AllPlantsCheck}) - 3), ${SpeciesList}))

Now it sounds like you may want to compare across multiple repeats? If so make a join() field for each repeat then make a master field that joins all of the joins together. Then check against that. (Confused yet )

Hope that helps. Works slick for me.

Re: Counting "unique" selections across repeats

RoryMcPherson — Mon, 02 Mar 2020 04:49:19 GMT

Hi Doug,

I'm trying to understand how to make this work, and am a bit confused about some of your explanation above. I'm mostly confused about which fields in your form are ${SpeciesList}, ${AllPlansCheck}, possibly a third text/calculate field(?), and what's going on in each one.

Here is what I have:

- A select_one species_list question in a repeat with the database field name 'plant_species'.

- Outside and below the repeat I have a unique_species_list text question, which in the calculation field I'm using join(', ', ${plant_species}) to add all the selected plants from the repeats to.

- On the same unique_species_list text question, I've added to the the constraint field: not(contains(substr(${unique_species_list}, 0, string-length(${unique_species_list})), ${plant_species})).

This is joining everything together, where I just want a list of distinct values. Also, choice list has several thousand species that are updated now and then, so I don't have any specific string-length codes. I've left this off the constraint, as I don't think it's required?

Oh, and would you be able to share this part of your form? It would be much easier for me I could see the survey sheet, but no worries if you can't share it.

Thanks in advance!

Re: Counting "unique" selections across repeats

DougBrowning — Mon, 02 Mar 2020 14:53:40 GMT

You must take at least something off of the string to check otherwise it will find itself. The add is instantaneous so the check would always fail.

Here are the lines. Hope it helps. You could remove the XXX check if you do not do Unknown codes. So just this -> not(contains(substr(${AllPlantsCheck}, 0, string-length(${AllPlantsCheck}) - 3), ${SpeciesList}))

type

name

label

hint

constraint

constraint_message

required

required_message

appearance

default

readonly

relevant

calculation

choice_filter

repeat_count

label::language1

hint::language1

media::audio

media::image

body::accuracyThreshold

bind::esri:fieldType

bind::esri:fieldLength

bind::esri:fieldAlias

body::esri:inputMask

text

AllPlantsCheck

All Species added so far

${SpeciesList}='' or contains(${SpeciesList}, "XXXX") or contains(${SpeciesList}, "UNK") or not(contains(substr(${AllPlantsCheck}, 0, string-length(${AllPlantsCheck}) - 3), ${SpeciesList}))

Species has already been added to the list

yes

join(", ", ${SpeciesList})

null

5000

Re: Counting "unique" selections across repeats

RoryMcPherson — Tue, 03 Mar 2020 03:41:14 GMT

Thanks for your quick response Doug. I gave this a go, but then found this post (Can I prevent a value from being submitted multiple times in a repeat question? ) and was able to use a JavaScript function to do this.

I appreciate the help though!

Re: Counting "unique" selections across repeats

DougBrowning — Tue, 03 Mar 2020 14:19:44 GMT

Does it work offline? I came up with my way a year ago before any javascript was available.

thanks

Re: Counting "unique" selections across repeats

RoryMcPherson — Wed, 04 Mar 2020 00:53:07 GMT

Great question! I had to check but seems to work fine.

Re: Counting "unique" selections across repeats

DougBrowning — Wed, 04 Mar 2020 16:24:32 GMT

Have you got it to work as a constraint? I can get a message to show like the example but when I try to use it as a constraint no go.

If I add to the note relevant it shows the note but when I add it to the constraint it still lets me keep going.

pulldata('@javascript', 'repeats.js', 'distinctCheck', ${SpeciesFound}) tried with and without the not()

If I add it to a join list constraint then it will not let me add anything.

I want it so that they cannot go on in the repeat with a dupe. That is what my way did. This gives a warning but still lets them do it.

Not sure what I am missing.

Thanks

Re: Counting "unique" selections across repeats

RoryMcPherson — Wed, 11 Mar 2020 01:03:28 GMT

The way I have it set up is using the distinct pulldata() function for species in a calculation for each associated plant type, which there aren't too many categories for. These are essentially just Boolean returns. I've then got a hidden species_check question using an if() calculation assessing if all the distinct checks return true. If any return false, as in a duplicate species has been selected for that plant type, the species_check returns a false; likewise, if all checks return true the check passes and returns true. I'm using this true/false return as a constraint on my select_one species in the repeat. This works for me as it triggers the constraint if they select the same species and then try to pass the repeat record.

Make sense? Hope that helps.

Re: Counting "unique" selections across repeats

JosephBusalacchi — Mon, 01 Jun 2020 15:51:08 GMT

Good Day Sir,

Any chance you can share the final spreadsheet showing the code you used to link all this together? I am working on a very similar project and would like to either limit a species from being entered twice or show a warning message if it is attempted. I get the basic portion of the Java code to work, it just shows the message all the time and I am not having any success in only showing the message when it is necessary.

Best,

Joey

Re: Counting "unique" selections across repeats

DougBrowning — Mon, 01 Jun 2020 17:33:04 GMT

I never got javascript to work how I want but my other way posted here https://community.esri.com/message/932144-re-can-i-prevent-a-value-from-being-submitted-multiple-times-in-a-repeat-question?commentID=932144&et=watches.email.thread#comment-932144

Does work for species.

Hope that helps.

Re: Counting "unique" selections across repeats

MarkWILSON_LLS — Fri, 29 Nov 2024 12:05:10 GMT

Hi there @RoryMcPherson , I realise this is an old thread but what youve described sounds like what Im trying to do.
I am allowing duplicates within my repeat (contains reserve name, area, ID plus a stack of field assessment data).
I want to add a flag or check field against the first instance a reserve is listed.

The second step in the main table would then be to Concat all the reserve names and Sum the Area of only the first instance records.

Are you able to post code of your desciption if you still had it?

thanks