Latest Contributions by RichardFairhurst

I am using ArcGIS Pro 3.2.  The Joins and Relates set up dialog performance is painfully slow to fill in the parameters.  For example, I have an active map with 35 layers/Standalone tables it takes up to 2 minutes to populate the parameter drop downs listing the available layers/tables and fields.  When the drop down options are available, the tool validation process virtually always resets all of my parameter choices back to the the original configuration when the tool opened approximately 4 times before it accepts my choices.  It has at times forced me to choose the values I want up to 25 times for up to 4 more minutes.  It seems to be stuck on validating the poorly constructed initial tool configuration it opened with which it can't satisfactorily complete before it takes any real input from me.  The validation is excessive and unresponsive and its effort to autocomplete the parameters for me are all counter productive. In the case of joins, I have had to wait up to 20 minutes for the joined table to refresh after completing the join on an SDE feature class and file geodatabase table with on 35K records in each and a one-to-one relationship.  All other open tableviews become disabled for table navigation and selection while the join I created is refreshing. Once the join of the 35 record tables was complete it took about 20 minutes to perform a simple selection query based on the join, even though the fields had attribute indexes.  This whole process and poor performance repeatedly occurred in multiple exclusive fresh sessions of Pro.   A field calculation also took 20 minutes. This was repeated with other SDE to file geodatabase joins with different record set sizes and all have performed similarly poorly at every step in the joins process.  Arcmap performed the entire set of tasks in under 4 minutes. Activating a relate and getting the results also seems to take considerably longer than it does in ArcMap regardless of the size of the selection that activated the relate. ... View more

The Attribute Transfer Tool Field Mapping can be set up to transfer the Source Shape field over to the Target Shape field.  In ArcGIS Pro 2.9 at least, when the Source Feature's polygon or polyline Shape contains true curves those true curves are not transferred to the Target Feature's Shape.  Instead, the Transfer Attribute tool transfers a Densified curve made up of straight line segments. Here is an example of a feature containing numerous true curves segments within its geometry.  If I use the Attribute Transfer tool to transfer the Shape of this feature back to itself, the feature is densified.  This should not be occurring. In ArcMap, true curves in the source geometry are transferred to the target feature's geometry when the Transfer Attributes tool is used.  Below is the same Attribute Transfer set up in ArcMap and the result of transferring the shape of the same feature with true curves back to itself. When I transfer the above features shape to another feature it retains any true curves that came from the source. ArcMap does not allow transfers from a feature back to itself.  This difference may be required in Pro since the Attribute Transfer tool Field Mapping also controls ArcGIS Pro's Copy/Paste behavior, which I have discussed in another post.  The Attribute Transfer tool in ArcGIS Pro needs to be made equivalent to the ArcMap version in its behavior of preserving True Curves.  This is in addition to numerous other behavior differences that I have described in other posts. ... View more

For many Polyline feature class data models a pair of fields or a set of four fields hold values that are dependent on the directional orientation of the line.  Examples of direction dependent pairs of fields include fields describing the left and right side of the line, such as the LEFT_FID and RIGHT_FID fields in the output of the Polygon to Line tool when the keep neighboring polygon information option is checked and Road Centerlines address fields holding city and zip code values for the left and right side of the line.  Field pairs can also be associated with the From and To ends of the line, which are direction dependent, such as fields for the X and Y values defining the ends of the lines.  An example of a set of four fields that are direction dependent upon both the left and right side of the line and the from and to ends of the line include Road Centerline fields for the Left From, Left To, Right From and Right To house number values of address ranges. A common problem arises when for any reason the directional orientation of the geometry of all or part of these lines needs to be flipped using the Flip Lines geoprocessing tool or the Flip modify feature editing tool and any of these direction dependent values needs to be flipped to match the new orientation.  There is no easy out of the box way to do that.  Using a geoprocessing approach to this problem involves creating either temporary fields or a temporary joinable copy of the table records to perform the flip/swap of these attributes, since the values of at least one of the fields needs to be temporarily preserved in a location that is separate from the pair of fields being flipped/swapped.  If a temporary field is created, three separate field calculations in a very specific order have to be performed to accomplish the flip/swap of just one pair of field values.  If a temporary joinable copy of records to be flipped is made, two calculations need to be done for each pair of fields.  And the selection of features that had their geometry flipped must be preserved at every step in order for the whole process to be done successfully. This entire problem is ideal for a tool that would include a field pair mapping parameter, like the ArcGIS Pro Calculate Geometry tool.  And numerous geoprocessing tools have been specifically designed to handle storage of data in memory or in temporary workspaces to remove the burden of managing these operations from the user.  This interface would let the user perform all value flip/swap operations in a single run and be so much easier to set up and execute than anything else currently offered. A sample field mapping for performing a Swap is shown below: Swap Field 1 Swap Field 2 LEFT_FID RIGHT_FID LEFT_FROM_NO RIGHT_TO_NO LEFT_TO_NO RIGHT_FROM_NO LEFT_CITY RIGHT_CITY LEFT_ZIP RIGHT ZIP FROM_X TO_X FROM_Y TO_Y   The tool would move every value held in one of these pairs of fields into either memory or a temporary workspace and then perform the swap of values between the two fields.  So, for example, for every record the LEFT_FID, LEFT_FROM_NO, LEFT_TO_NO, LEFT_CITY, LEFT_ZIP, FROM_X and FROM_Y values would first be temporarily stored in memory. Then move the RIGHT_FID value to the LEFT_FID field, the RIGHT_TO_NO to the LEFT_FROM_NO field, the RIGHT_FROM_NO value to the LEFT_TO_NO field, the RIGHT_CITY value to the LEFT_CITY field, the RIGHT_ZIP value to the LEFT_ZIP field, the TO_X value to the FROM_X field and the TO_Y value to the FROM_Y field. Next, move the LEFT_FID memory value to the RIGHT_FID field, the LEFT_FROM_NO memory value to the RIGHT_TO_NO field, the LEFT_TO_NO memory value to the RIGHT_FROM_NO field, the LEFT_CITY memory value to the RIGHT_CITY field, the LEFT_ZIP memory value to the RIGHT_ZIP field, the FROM_X memory value to the TO_X field and the FROM_Y memory value to the TO_Y field. Finally, perform any memory and temporary workspace clean-up for the user. This whole set of operations being done one step at a time by the user is prone to errors, since it involves a minimum of either 16 steps or 21 steps all of which involve working with very similar field names that can easily be confused or difficult for the user to keep track of.  Even using the Calculate Fields tool, which has a field mapping, it can be confusing, because the operation involves a minimum of three separate steps that each involve numerous field with very similar names that have to be performed in an exact order. The interface set up above is much easier to understand than any manual workflow I could describe and once the tool is developed and debugged these operations become standardized and will be performed infallibly and with easy repeatability once the user configures the field map correctly for a given Polyline feature class.  This tool would naturally follow or precede the use of the Flip Lines geoprocessing tool as a separate step for the same feature selection.  Having this tool in combination with the Flip tools would allow the user to make simple modifications to the field map if they only want to swap a subset of these attributes because they know that the other attributes are oriented correctly with the flipped geometry already and may have been the cause of the geometry flip in the first place. ... View more

Being able to retain foreign key values in any extract of data that the user wants to create that match the source primary key values is fundamental to most data administration workflows.  Extractions of records from a source that relies on its OBJECTIDs and/or its GLOBALIDs as its primary key in relationship to other records should not be relationally destroyed relative to the source that they came from if the user has an explicit requirement to preserve those values as a foreign key in their extracted output.  However, this kind of relational destruction is currently always happening whenever the Feature Class to Feature Class and Table to Table tools are used, which are the default tools for creating extractions in ArcGIS Pro. The Feature Class to Feature Class and Table to Table tools (and possibly other tools) need to add two options that would allow the user to create new foreign key fields that preserve the OBJECTID and/or GLOBALID values of the source feature class/table in the output.  If the user checks either of these options a control in the tool should suggest a default Field Name for each which the user could override.  Validation should ensure that the source contains OBJECTID/GLOBALID fields before enabling these options, and that the user definable field names are valid and not already in the source feature class/table or duplicated if the user checks both options.  The source OBJECTID values would be preserved in a Long field and the source GLOBALID values would be preserved in a GUID field. The use case for this is when the only field(s) that contain a unique value for each of the source features or rows is the OBJECTID and/or the GLOBALID.  Often the source table is not owned by the user and cannot be modified to store these values into a Long and/or GUID field in advance.  Using either of these tools overwrites the source OBJECTID and GLOBALID values with new values in the output, especially when the user only wants to extract a subset of records.  Having to copy the source data into a workspace the user can modify to add these fields in advance is very time consuming when the source contains large numbers of features and/or rows and they have to repeatedly copy the source every time they need a new extraction if they have any reason to believe that the source contains new records that they want.  This Copy workaround is especially burdensome to the user who only wants a small subset of the source features or rows and not an entire copy to apply the fundamentals of relational database design with their output.  There are also many instances when the Copy function is not supported for even doing this workaround when the source feature class is part of a topology within a feature dataset. Incorporating the preservation of the OBJECTIDs in a static foreign key field in the output of these two tools is absolutely possible, as demonstrated by numerous tools like Spatial Join, Union, Intersect, Polygon to Line, Multipart to Singlepart, etc. all of which preserve a foreign key FID field in their output.  Expending this kind of foreign key preservation behavior to also allow for the preservation GLOBALID values in a GUID foreign key output field should not be significantly more difficult.  This would be a great enhancement and time saver for users who need extractions from Parcel Fabric data sources and any other advanced feature dataset capabilities that are built around GLOBALIDs as primary keys for relationships. And adding a GLOBALID to GUID option to many other tools that currently by default preserve OBJECTIDs in a foreign key FID field would be great as well as a time saver.  But that is less critical, since at least those outputs already preserve usable foreign key values from the source OBJECTIDs in their outputs by default.  Additionally some of those tools support field mapping, which may allow the user to accomplish the GLOBALID value preservation themselves.  But the Feature Class to Feature Class and Table to Table tools currently do not offer any option that allows the user to preserve any usable foreign key values in their output relative to any sources that rely on OBJECTIDs or GLOBALIDs as their sole unique primary keys. If the existing Feature Class to Feature Class and Table to Table tools should not be modified to add this option because they are accessible to all license levels and intentionally provide limited functionality as a result, a new pair of Conversion tools called Feature Class to Feature Class with FIDs and Table to Table with FIDs should be added for Advanced License users at minimum.  These two new tools would definitely be a selling point for many Basic and Standard users to consider upgrading to an Advanced license, as well as two very befitting and useful additions for truly Advanced users. ... View more