Solved

Shapefile Writer Issue (missing .prj file)

  • 11 November 2022
  • 8 replies
  • 8 views

I'm processing a point feature class (21 million approx points) I'm fanning them out into shapefiles, smaller than the 2G limit.

 

It produces shapefiles, missing the .prj file. Any thoughts on why ?

icon

Best answer by joneszero 29 November 2022, 01:09

View original

8 replies

Userlevel 2
Badge +13

Hello @joneszero​, thanks for posting! This seems a bit odd/interesting. Shapefile writer documentation defines .prj as "Projection information for spatial referencing geometry to locations. If present, will be used by FME to georeference geometry". I noticed you're using a custom coordinate system, if you use a standard coordinate system (eg. one that is shipped with FME), are PRJ's written properly? I suspect FME should be able to write the PRJ's even when a custom coordinate system is being used, but curious.

 

Alternatively, what version of FME are you currently using? Does the same behaviour occur in the latest FME 2022.2 official build? Would you be able to try please? Also, if you're able to share a sample input or the custom coordinate system? Happy to help, Kailin.

Hi,

Thanks for your help.

Yes, I'm using a custom GCS (WGS1984_NSIDC_EASE-Grid 2.0_Global).

I'm currently using FME(R) 2021.2.0.1 (20211112 - Build 21789 - WIN64). I'll try it on the latest build, as well as using a standard GCS, get back to you.

I'm processing AIS points (maritime vessel tracking data), see sample data attached.

 

 

 

Here's the custom GCS.

Userlevel 2
Badge +13

Here's the custom GCS.

Hello @joneszero​, thanks for your patience, apologies for the delay in response. I wasn't able to reproduce the issue you're experiencing. Is the provided 'Temp' data an input or output of the ProcessAIS.fmw workspace?

 

I notice you're using a File Geodatabase Reader in the workspace. Would you be able to share an input sample as a geodatabase instead of a shapefile? If you're unable to share it on the FME Community, consider submitting a support case with us.

 

Are you able to provide a screenshot of the folder containing the output shapefiles (with the .prj's missing)? Otherwise, can you reduce the number of files being written by FME, to see if this has an effect on the .prjs being created or not. Happy to help, Kailin.

Hello, the "temp" file is an input data source.

 

See attached a sample of the input data (geodatabase).

 

I'll that screen shot through shortly.

 

thanks for your help

 

Well, it seems I've found a work around.

In ArcGIS Pro, I exported the source data (FGB, point feature class) into a newly created FGB. Ran this identical but "newer" dataset through the work bench, and the output shapefiles are complete and readable.

I'm guessing it must be issue with the version of ArcGIS Pro the FGB was made in, and the version that I'm using ?

Jeremy

Userlevel 2
Badge +13

Well, it seems I've found a work around.

In ArcGIS Pro, I exported the source data (FGB, point feature class) into a newly created FGB. Ran this identical but "newer" dataset through the work bench, and the output shapefiles are complete and readable.

I'm guessing it must be issue with the version of ArcGIS Pro the FGB was made in, and the version that I'm using ?

Jeremy

Hello Jeremy @joneszero​, so interesting/odd. I still haven't had any luck reproducing the issue. Would you be able to share what version of ArcGIS Pro the original FDGB was produced in? Alternatively, would you be able to share a copy of the original FGDB? If we have a copy of the original dataset, we may be able to poke around more / shed some light as to what is going on. Regardless, I appreciate you bringing this to our attention. Happy to help, Kailin.

Hello Jeremy @joneszero​, so interesting/odd. I still haven't had any luck reproducing the issue. Would you be able to share what version of ArcGIS Pro the original FDGB was produced in? Alternatively, would you be able to share a copy of the original FGDB? If we have a copy of the original dataset, we may be able to poke around more / shed some light as to what is going on. Regardless, I appreciate you bringing this to our attention. Happy to help, Kailin.

Hi Kailin,

A copy of the original FGDB is too large, greater than the 40MB attachment limit. I'm not aware of what version of ArcGIS Pro the original FGDB was produced in.

Some more background info ...

The workspace separates the shapefiles written according to the maximum number of records for a shapefile (500000). I found that when I would run the workspace on just a sample of the data the shapefiles written would be fine (e.g. multiple shp weren't produced, just one). However when it was ran on the whole dataset, the shapefiles written (at max 500000 record size) were incomplete, or written without .prj files.

Thanks for your investigations on this.

Jeremy

Reply