Solved

Problem with swedish characters when writing to ESRI geodatabase

  • 12 January 2016
  • 4 replies
  • 7 views

Badge

I have a workspace where I read a shape file, clip the data and write it to a ESRI geodatabase with the

 

Esri Geodatabase (ArcSDE Geodb) reader. The attributes in the shape file contains swedish caracters. The characters is right when I read the shape file in ArcGIS but when I write the data to the geodatabase the swedish characters is replaced. How can I handle that? I have try to change character encoding on the shape file writer to iso-8859-1 (which contain swedish characters).

I'm attach print screens from ArcGIS when reading the shape file (as it should look like) and from ArcGIS when reading the geodatabase.

icon

Best answer by david_r 12 January 2016, 13:53

View original

4 replies

Userlevel 4

It's always difficult to debug these issues remotely, but I agree it is probably an encoding issue.

Two things to try:

  • On the shape file reader, try also using the "Windows Latin1 ANSI (1252)" encoding
  • Just before the writer, insert a Logger or an Inspector and look at the characters and the encoding that is reported by FME. Example:

David

Badge

I try to use Windows Latin1 ANSI (1252) encoding. The logger I insert before the writer gives me

Attribute(encoded: windows-1252): `Lan' has value `™sterg”tlands l„n'

FME report the same encoding but i get the same attribute value as before.

Userlevel 4

I try to use Windows Latin1 ANSI (1252) encoding. The logger I insert before the writer gives me

Attribute(encoded: windows-1252): `Lan' has value `™sterg”tlands l„n'

FME report the same encoding but i get the same attribute value as before.

What happens if you set the Shape file reader encoding to Unicode UTF-8?

Basically, you need to experiment with the reader encodings until it looks right in the Logger.

In most cases, it is either Latin1 or Unicode. If your dataset is really old it might be worth trying "DOS Nordic (ibm865)" as well.

Badge

What happens if you set the Shape file reader encoding to Unicode UTF-8?

Basically, you need to experiment with the reader encodings until it looks right in the Logger.

In most cases, it is either Latin1 or Unicode. If your dataset is really old it might be worth trying "DOS Nordic (ibm865)" as well.

Thank you for your help! It work's fine with DOS Nordic.

Reply