Question

change input data (same file structure)

8 years ago
27 January 2016
7 replies
1 view

jamiefinney
22 replies

Hi i've only just started using FME (2016) so this is likely a very simple problem.

i've created a workspace that takes a series of MI tab files, and re-maps some attributes to a writer. it works fine. now that its finished i want to run the same tool again with different data (same structures). what is the best way of doing this?

i've tried right clicking on the reader > edit 'NAME [MAPINFO]' parameters and changing the data there and that doesn't seem to work when i run the tool. nether does reader menu > update feature types.

Data_1 Col_A Col_C

Data_2 Col_C Col_B

Data_3 Col_B Col_A

Output Col_A Col_B Col_C

7 replies

Userlevel 4

Have you tried simply changing the path to the new tab file under the published parameters? You will find the published parameters in the Navigator on the left hand, some place under the readers and writers.

David

@david_r thanks for your reply it looks as though this does the same thing as right clicking on the reader > edit 'NAME [MAPINFO]' parameters as these have updated to the new values in the navigator also.

the file structure of the new data is identical to the old one. i'm running with full inspection and nothing is coming out of the reader ports. i get an "unexpected input dialog" stating that the input data has changed (which it obvious has) not sure how to get around this.

Userlevel 5

+25

redgeographics
Influencer
3345 replies
8 years ago
27 January 2016

Mapinfo is a folder-based format, so each Mapinfo file inside a folder corresponds to a feature type and they're being matched by name. It sounds like there might be a discrepancy there. E.g. you set up your process to read file1.tab and file2.tab, but now have changed that to file3.tab and file4.tab and you're getting unexpected input data.

To solve this, you can go into the properties of your reader feature types and either rename them to match your input data or set up a merge filter (either by wildcard or regex)

so i tried the option to merge using a * and nothing went through. what does seem to work however is deleting the reader and replacing it with a new one (pointed at the new dataset). (not ideal us my model has 7 input readers) and using an attribute manager rather than the manual use of arrows to link datasets to writer.