Update from AGP 3.1.3 to 3.2 causes "Points to Line" tool to malfunction (cause found as the thread went on)

[MaxRingler1]

I wanted to use the "Points to Line" tool in AGP 3.2 to let it do what its name says - group points in a shapefile based on a text field and then connect them, based on a date field, with all environment values on default. However, when running, the tool mostly gives me zero-length lines for every point (there a a view groups with only one enty, but the far majority have 2-20 entries). A few lines, always only connecting 2 points are drawn on the map. However, when I select them on the map, no selection is made in the attribute table and when I select them in the attribute table, no selection is made on the map. The behaviour is idential with the shapefile having projected and geographic coordinates.

Using the same tool in ArcMap 10.8 works line a charm, on the identical projected and geographic shapefiles - indicating to me that the source shapefile is not corrupted.

Does anyone have an idea or suggestion what might be the issue here?

Cheers

Max Ringler
University of Bern

[DanPatterson]

what does the table entry look like? and the dialog parameter selections?

[MaxRingler1]

For the "Line Filed (grouping) I use the field "INDIVIDUAL" which is a 9-character text string in the form of "i20-m-123", for the "Sort Field" (ordering), I use the filed "DATE_" which is of the date-data type in the form of 01.01.2023. I tried to create alternative grouping and sorting fields (ID & ORD) with manually entered short integer values, corresponding to INDIVIDUAL and DATE_, but this only yielded the identical corrupted results as described before.

In the "Details" the geoprocessing tool throws a ton of error messages - all related to the claimed singletons in the data (

WARNING 001059: Not enough points to create a line for value i23-m-010.

However, all these values indeed have multiple entries with the same value.

[DanPatterson]

Your Individual field and the warning means that you need at least 2 observations to form a line.

Have a look at the table (perhaps just sort it and weed out the rows that dont have duplicates, those are the singletons)

[MaxRingler1]

I understand the error message. However, as I said 99 percent of the 323 entries are *not* singletons, but still I get 317 error messages and the algorithm claims to only sees 3 pairs. This happens when grouping is based on the original INDIVIDUAL text field and on the secondary ID integer field 

[MaxRingler1]

It's more mysterious - I now tried the same thing with the same project and the same shapefile on another computer, and there it works fine and gives the expected result!

The computer with issues is a newly setup Win11 PC with a clean install of AGP 3.2 2 weeks ago. It has a parallel install of ArcMap 10.8.

The computer without issues is a ~10 year running Win10 PC, running APG 3.1.3 (!), also with a parallel install of ArcMap 10.8.

To give it a try, I now updated this machine from AGP 3.1.3 to AGP 3.2 - voilà, now the previously good machine causes the same erronous output when running "Points to Line" in APG 3.2. 

[MaxRingler1]

As "Points to Line" is a script-tool - might this issue be due to a how AGP interacts with Python, or the Python version installed?

[DanPatterson]

At this stage, I would open an issue with Tech Support

[MaxRingler1]

Done...thanks for staying with me 😉

There is a probably related issue already reported: BUG-000163006 for ArcGIS Pro (esri.com)