topic Re: Learning about SRIDs (and constructing new geometries) in ArcGIS Pro Questions

Learning about SRIDs (and constructing geometries via SQL)

Bud — Fri, 03 Jun 2022 22:25:03 GMT

Edit: Mods, I suppose this post could be moved to the Data Management community (or whatever's appropriate).

What resources do you recommend for learning about SRIDs? SRIDs have always been a gap for me.

For example, when working with a FC using SQL, it's not clear to me if/when I would need to change the SRID for the output geometries.

If I extract points from a line FC via an SQL query:
When constructing the resulting points using SQL, should I use the original SRID from the lines, or do I need a new SRID since the geometry type has changed? (from lines to points)

select 
    objectid,
    --Side note: If anyone's wondering, this is why I convert to/from WKB: https://support.esri.com/en/Technical-Article/000011333
    sde.st_geomfromwkb(sde.st_asbinary(sde.st_startpoint(sde.st_geometryn(shape,1))),sde.st_srid(shape)) points_shape
    --                                                                                   🡅
    --Question: When constructing the resulting points, should I use the original SRID from the lines, 
    --    or do I need a new SRID since the geometry type has changed? (from lines to points)
from
    lines    --Note: The original SRID of the lines is 300051.

Or, if I remove the M-dimension from lines via an SQL query:
Do I need to use a different SRID when constructing the non-M-enabled geometries via SQL?

Thanks.

Re: Learning about SRIDs (and constructing new geometries)

Anonymous User — Fri, 03 Jun 2022 17:36:54 GMT

Just curious, where does SRID 300051 come from? A quick search doesn't show anything. Is that an implementation detail of the database you're using?

Re: Learning about SRIDs (and constructing geometries via SQL)

George_Thompson — Sat, 04 Jun 2022 13:31:10 GMT

For SRID info, take a look here: https://desktop.arcgis.com/en/arcmap/latest/manage-data/using-sql-with-gdbs/what-is-an-srid.htm especially for the 300### ones

Re: Learning about SRIDs (and constructing new geometries)

Bud — Fri, 03 Jun 2022 19:19:15 GMT

That SRID 300051 comes from ArcMap. I'm not sure what the reasoning is behind it:

Re: Learning about SRIDs (and constructing new geometries)

Anonymous User — Fri, 03 Jun 2022 21:14:02 GMT

After reading @George_Thompson's link, I think a general answer might be ok here. TLDR Just use the same SRID. I don't work out of a database in this manner so this answer is just from principle.

Assumptions:

1. Your underlying coordinate geometry is in the correct SRID. The projection or geographic coordinate system has been defined correctly.

2. You do not need to convert or transform geometry to accomplish your analysis.

3. Edit to add an assumption: All records in the table have the same SRID. Apparently records in a table don't have to share an SRID.

If these assumptions are true and you are changing the geometry type, just use the same SRID. If the feature geometry's coordinate system is unknown or incorrect, you would need to (re)Define it. If you need to use another SRID for an analytic purpose, you would need to Project it.

From the docs:

Unknown coordinate system with an ST_Geometry, SDEBINARY, PostGIS geometry, SDO_Geometry, or SQL Server geometry column

An SRID greater than 300,000 is added to the geodatabase system table.

Definitely wondering what's going on here. What does the XY Coordinate tab say?

Re: Learning about SRIDs (and constructing new geometries)

Bud — Fri, 03 Jun 2022 22:27:36 GMT

Thanks. That helps.

Here's the XY Coordinate System tab of the original FC:

And here's the Domain, Resolution and Tolerance tab:

Re: Learning about SRIDs (and constructing new geometries)

Bud — Fri, 03 Jun 2022 22:28:12 GMT

I did a test just now where I created a new FC from scratch using Catalog (using all the default settings).

The SRID is 26917. So that seems more normal, yes?

Maybe there was just something something funky going on with my original FC.

Re: Learning about SRIDs (and constructing new geometries)

Anonymous User — Fri, 03 Jun 2022 22:26:46 GMT

At least that new "Default" SRID corresponds to something more concrete in the registry.

Since you asked about learning, here are the basic questions I would want to answer for myself before proceeding.

- Is NAD 83 UTM 17N the expected coordinate system for this data?
- Are the real world features within the NAD 83 UTM 17N zone?

- Does an SRID 300051 feature class behave the same as a NAD 83 UTM 17N feature class during the same analytic process (e.g. your SQL query and comparing results)?

If the answers are all yes, that's good preliminary confidence to continue using the same SRID. Worth noting that best practice is to project all data into a common CRS.

Re: Learning about SRIDs (and constructing geometries via SQL)

Bud — Sat, 04 Jun 2022 17:39:13 GMT

@George_Thompson Thanks!

Side note: Is there a chance there's a small typo in your comment? Should "3000###" be "300###"?
That's obviously not a big deal. I just thought there's a chance it could cause confusion for someone else reading this post.

Re: Learning about SRIDs (and constructing new geometries)

George_Thompson — Sat, 04 Jun 2022 13:31:28 GMT

Thanks for catching the extra "0". Post updated.