Filter gallery or collection to only show entries with duplicate values

[Anonymous]

Good day to everybody,

Sorry if this was already solved somewhere else, but all the googling from the last two days didn't show any helpful results.

Quick intro to my needs:

- I have a document library for invoices with some folders, and in all those folders are PDF files with all sorts of meta data, one of them is the invoice number

- it must be possible to add the same invoice number again, just in case it happens that we get it from different sources with same invoice number, so this can't become a unique value

Because it happened / happens, that the same invoice was handled by different persons, it was also stored in the library >1.

Now I was trying to build a collection / gallery that is doing these actions:

- Get all invoice numbers from all folders

- remove / hide all invoice numbers which only appear once

- show only the duplicate entries

--> pretty much the oposite of distinct I guess

Currently I got this:

Collect all data from the data source "Dokumente" except if the invoice number (RG_x002d_Nummer) is not filled yet

Collect(RechNum1; Filter(Dokumente; If(IsBlank(RG_x002d_Nummer); false; true)))

From this collection I build a gallery which only shows distinct values for invoice number and CountRows AllItems

Search(Distinct(Filter(SortByColumns(RechNum1; "RG_x002d_Nummer"; Ascending); If(IsBlank("RG_x002d_Nummer"); false; true)); RG_x002d_Nummer); SearchInput_99.Text; "Result")

In order to see items which have a duplicate invoice number, I made a CountIf where I count the amount of similar invoice numbers per item

CountIf(RechNum1; RG_x002d_Nummer = 'lbl_RG-Nummer_1'.Text; true)

I then fill those items with a reddish color in order to see them better

If(CountIf(RechNum1; RG_x002d_Nummer = 'lbl_RG-Nummer_1'.Text; true) = 1; LightGreen; LightCoral)

Next to this I have another gallery listing all items without "distinct", if the CountRows of AllItems of those two galleries are different, I already know that there must be duplicates.

Search(Filter(Search(SortByColumns(RechNum1; "RG_x002d_Nummer"; Ascending); SearchInput_99.Text; "RG_x002d_Nummer"); If(IsBlank(RG_x002d_Nummer); false; true)); SearchInput_99.Text; "RG_x002d_Nummer")

This is the result from what I explained above, the left gallery has the "distinct" filter, the right is without it.

With only 226 invoices its ok to scroll, but if I want to crosscheck with our archive, this list extends to above 2000 items and scrolling is not an option anymore.

So what I am trying to accomplish is building a gallery where only those invoice numbers will be listed, which are duplicates.

I hope this was not too confusing, would be very thankful if anyone would have a hint leading me into the right direction 

Thanks in advance!

Cheers

BJ

[RandyHayes]

@Anonymous 

You might want to look at the GroupBy function for this rather than the logic you have now.

You can replace it all with this formula:

AddColumns(GroupBy(colInvoices, "invoiceNumber", "invoices"), "docCount", CountRows(invoices))

NOW...your actual problem is the number of documents that you have being beyond 2000.

To have accuracy in your App, you will need to Collect all the invoice numbers into a local collection in multiple collects to have the total list of invoices locally.  Once you have that, the above will be exactly the same.

I hope that helps you some. 

[Anonymous]

@RandyHayesthanks a lot for your quick response.

Besides the issue with above 2k items, am I correct that in your example you GroupBy similar invoice numbers, but you still show all invoice numbers in the gallery?

My main issue is that I don't want to list the ones, that don't have duplicate entries.

Forgive me if I am wrong, but according to your left image it seems as you show all items in the gallery.

Thanks for clarifying 

[RandyHayes · Accepted solution]

@Anonymous 

No problem.  In that case, just filter the whole thing.  Formula now is:

Filter(
    AddColumns(
        GroupBy(colInvoices, "invoiceNumber", "invoices"), 
    "docCount", CountRows(invoices)), 
docCount>1)

---

Besides the issue with above 2k items, am I correct that in your example you GroupBy similar invoice numbers, but you still show all invoice numbers in the gallery?  When you GroupBy, the column that you group by it is the same as doing a distinct.  If there are 500 invoices.  300 are distinct names, then your GroupBy results will have 300 records. Within those records are the filtered records that matched that name. If you were to add up all the individual records within in the grouped records (in this case in the invoices column) you would have 500.

My main issue is that I don't want to list the ones, that don't have duplicate entries.  See formula change above.

Forgive me if I am wrong, but according to your left image it seems as you show all items in the gallery. Yes, that was correct.  In my original sample, I was demonstrating the use of the GroupBy function to achieve the results that you had in several collections and formulas.  I glossed over the fact that you wanted to filter that by only ones that were more than 1.  BUT...the above correction will give you that.

---

See how that all works out for you now. AND, of course, this still doesn't address the 2k limit.

[Anonymous]

Good morning Randy, Thanks so much for the great help!

I actually stayed with my way of collecting the data, just because I knew what I did there 

But from there I made a second collection with your Filter into a ClearCollect where I grouped them and added the column for the count of the elements.

So OnVisible this happens first:

Collect(RechNum1; Filter(Dokumente; If(IsBlank(RG_x002d_Nummer); false; true)));; ClearCollect(RechNum2; AddColumns(GroupBy(RechNum1; "RG_x002d_Nummer"; "RG_Nums"); "RG_Count"; CountRows(RG_Nums)))

And my gallery, where I only show the items with duplicate values, looks like this:

Search(Filter(RechNum2; RG_Count >1); SearchInput_102.Text; "RG_x002d_Nummer")

What I realized now thou, some invoice numbers have a blank at the end, so know I have to get rid of the blank space at the end of the invoice number, but I guess I will make a new post for this in case I can't figure it out.

Once again thanks!

Cheers

BJ

[RandyHayes]

@Anonymous 

So if extra spaces are an issue, you can use the Trim function to clean that up.  

Using the formula I provided before, your Gallery Items property would be:

Filter(
    AddColumns(
        GroupBy(
           AddColumns(colInvoices, "cleanInvoiceNumber", Trim(invoiceNumber)), 
        "cleanInvoiceNumber", "invoices"), 
    "docCount", CountRows(invoices)), 
docCount>1)

[Anonymous]

Thanks for this also, but in my "build" I seem to have issues implementing the Trim("RG_x002d_Nummer") function.

Here's the code I tried to use it in, as this creates my final collection for the gallery which only shows the duplicate items:

ClearCollect(RechNumAll2; AddColumns(GroupBy(RechNumAll; "RG_x002d_Nummer"; "RG_Nums"); "RG_Count"; CountRows(RG_Nums)))

Do I use the Trim("RG_x002d_Nummer") while creating the collection or do I use it while filtering the gallery?

To my understanding the Trim("RG_x002d_Nummer") must happen before the CountRows, otherwise the trimmed items wont be counted.

Maybe, as we would say in Germany, I am just standing on the hose 

[RandyHayes]

@Anonymous 

No, the trim has nothing to do with the counting, it has everything to do with the grouping.

If you're going to use that in a collection, then this Formula will be what you want (I see you're using ; rather than , so I'll adjust, and I'll add your names):

ClearCollect(RechNumAll2;
     AddColumns(
        GroupBy(
           AddColumns(Dokumente; "sauberRGnummer"; Trim(RG_x002d_Nummer)); 
        "sauberRGnummer"; "rechnungen");
    "zählen"; CountRows(rechnungen))

 What this does exactly, is to first create a column on the Dokumente source for sauberRGnummer which is a clean invoice number without extra spaces (the Trim function on the RG_x002d_nummer cleans it up).

Then it groups by the clean invoice number sauberRGnummer.  You now have a table with two columns - sauberRGnummer (distinct list of clean invoice numbers) and rechnungen (a table of all the invoices that match that number).

Then we wrap all of that in another AddColumns statement that will create another column in the above table which will be the count of the rechnungen table in each row.  This will be your document count.