Interpreting a text file through sequential reading

Hi,

 

 

look into the multiple feature support of the AttributeCreator (http://docs.safe.com/fme/html/FME_Transformers/Default.htm#Transformers/attributecreator.htm), which can reference previous or subsequent features to join your TIMESTAMP features with the preceeding DATESTAMP features.

 

 

David

 

 

Hi David

 

 

Thanks - that's useful, I wasn't aware of that. Unfortunately, I still don't know how many features to tell the AttributeCreator to look backwards for, as this varies depending on the number of licence issues on a given day. Is there a way that I can test for a certain condition within this e.g. contains "datestamp"?

 

 

Thanks, Riley

Hi,

 

 

You can use a txt reader, so it reads textline_data.

 

Expose line_number (if u intend to write it to a textfile again)

 

Connect a stringsearcher with a regexp to find DATESTAMP: (.*)

 

Conne the found port to a variableSetter: variablevalue = matched_parts{0} (this stream does not connect further to anyting, it ends there)

 

 

 

Connect  a variableretriever to the "not found" port and retrieve the variable u set datestamp to and assign it to a attribute (for instance "d_Stamp").

 

 Then connect to a  stringsearcher and serach  for TIMESTAMP: (.*) and assign it (matched_parts{0} ) to a attribute (for instance t_stamp.

 

Now you should have the textfile read an all t_stamps belonging to a d_stamp are related.

 

 

(this is because of sequential reading and the fact that the textfile is ordered)

 

 

 

Gio - this has solved for me. Thank you very much to both of you for quick responses.