Latest Contributions by VinceAngelo

You need to put a WHERE constraint on the QueryLayer definition to restrict by either GeometryType(geomcol) or ST_GeometryType(geomcol), then specify the desired type in the layer definition  Those two functions return different values (e.g. "POINT" and "ST_POINT"), and you may need to use an IN operator, because multi-part linestring and multi-part polygon geometeries are represented as different types in PostGIS. You may want to use ST_Dimension(geomcol) instead, since that just sorts POINT, LINESTRING, and POLYGON data into 0, 1, or 2 (but be careful, because MULTIPOINT also returns 0, and it's a different type as well, though not as common).  In a perfect world, the various geometry types wouldn't be smushed together like that (it's an antipattern, but some folks love to go down that route).  You'd get better performance on large tables (>100k features) if the dimensionality were also an attribute property capable of being indexed (e.g. 'WALL', 'WELL','PATIO') though I suppose a covering index on one of the differentiator functions could be effective (but only if you can get the data owners to construct that index). - V ... View more

I've got a couple of read-only RDS replicas exposing PostGIS geodata via Query Layers. Works great. Selection sets would likely be a problem from an EGDB, since the database is read-only. EGDB replication would likely be better than an RDBMS replica for true geodatabase functionality. - V ... View more

Yes, this is unfortunately true. - V ... View more

According to Stack Exchange, you want Latin1_General_100_..._UTF8 - V ... View more

No, latin1 is NOT UTF-8, it's latin1. You need to use a UTF-8 character set. - V ... View more

I've been UTF-8 compliant on my databases for 22+ years now. There really is no reason to go out of your way to make international names like "Côte d'Ivoire" give your front-end developers nightmares. - V ... View more

I've had good experiences with leveraging an SHA-1 hash of the non-key important data*, storing the hashes with the actual primary key in a parallel table.  Due to the nature of the data, I've sometimes converted the data to only a few significant digits, then hashed that (to enable "fuzzy matching", when 3.71 -> 3.72 or 3.68 is not significant, but 8.23 is). Recently I've stored the geometry and non-geometry hashes separately, and have even been able to match rows without a useful rowid key by inverting the hash dictionary spanning tens of millions of features, so that hashes could resolve feature matching at 85-97% accuracy. Unfortunately, I can't release my code without  approval. And in order to make this work , you need to  be really comfortable with endian integer and floating-pint storage representations. I can say the core code is a set of encoder functions (one for each supported type), which is stored as a bytearray sequence, that is fed into a core Python hashing function. Use of SHA-1 can generate safety warnings, but using it for hashing (not cryptographic) purposes is perfectly safe. - V --- *Unimportant data would include a URL that incorporates the primary key and another column, which could be recomputed (or changed at will) ... View more

You need to go back and review the documentation on versioning. Versions stay forever, until deleted. They also pin subsequent edits in the A/D tables, making the geodatabase less efficient. This is not something which should be ignored. If that initial set of edits in that version was just to prove it was possible, and never intended to be merged into the table, the version should be removed, and the state tree compressed, to eliminate the exact problems you're having now. Versioned editing is a process, with primary use cases designed to incorporate the edits as quickly and efficiently as possible. Coloring outside the lines of best practice can stress the implementation and result in suboptimal performance. - V ... View more

One simple heuristic: Use Create Fishnet (Data Management) to partition the space, then iterate the features, using them to Select Layer By Location, and compare the number of selected features. If fishnet cell size is, say, 20-50% larger than the usual subdivision area, missing subdivisions will stand out in the feature counts. From a computer vision standpoint: Symbolize the features with exceptionally wide line symbols, in two different primary or secondary colors with 50% transparency, then make a series of tiled images, and look for images (areas) where one color predominates. - V ... View more

Doug -- Views are, well, views. The geodatabase doesn't have the XML metadata from the original table(s)' columns, so it can't know what the aliases were. So it's not a matter of "dropping" them so much as never having had them. And if the view can't participate in geodatabase behaviors, it doesn't get the XML that permits aliases to exist. Note that nothing is ever "in the SDE". It's always been in the database, accessed through what was last called ArcSDE (but nowadays, it's all just ArcObjects). SDE is long gone. "There is no Dana, only Zuul." - V  ... View more

Please edit the question to place the code in a code block (click on "..." then "</>" in the second row of icons). Reading Python code without indention is nearly impossible. It's also nearly impossible to debug incomplete code (especially when there are TWO SearchCursor flavors [and arcpy.SearchCursor is deprecated, and should not be used in code written since 2013]). But you should know that nesting cursors is an antipattern, and the likely cause of your error. - V ... View more

The PostgreSQL team made a significant change at PG 12 that eliminated a hidden column that Esri was using as the primary key on the objectid allocation algorithm. I'm not sure you can make the leap from PG10 to PG15 just by doing a backup/restore, since all your "i-tables" are now invalid. If the tables aren't versioned, you might be better off exporting through file geodatabase, or just copying from the PG10 instance to the PG15 one (you can run both on the same box using different ports). Tech Support likely has better tools to work out the problem here. - Vince ... View more

You have corrupted geodatabase metadata. TLDR solution: Rename the table to a temporary name Create a dummy table (just objectid is enough) with the "existing" name Use ArcObjects (Pro UI or ArcPy) to delete the dummy table Rename the temp name back to original Register the no-longer "existing" table - V ... View more

Management Studio is a client to the SQL Server database, just like Pro, so it doesn't really have any role in enterprise geodatabase compatibility.  You don't state the database server version in use, so it's hard to determine compatibility, but that message seems to come from the database drivers, not from Esri components. Note that it's recommended to upgrade your clients to the intended Enterprise version before upgrading the ArcGIS Server/Portal components, so you really ought to be running ArcGIS Pro 3.4 with Enterprise 11.4. The software compatibility matrix is available in this FAQ topic. - V ... View more

Well, if you had used the sde schema, not sdo, and didn't register any user tables, then dropping the sde schema (including contents) could have been enough.  And if you can identify all the dbo tables that were added (most have 'sde_' or 'gdb_' prefixes), you might be able to pull this off (again, provided you didn't register any tables). Once you've registered a table with the geodatabase (with either sde or sdo model), the cleanup becomes significantly more complicated, and falling back to the last backup (Ryan's suggestion) makes good sense. - V ... View more