Hi,
I have a process flow which reads daily txt files in as follows;
List of Files -> Iterate List of Files
In the Iterate Loop I have
Variable Based File Reader (a) -> [some actions] -> Concatenate (so now a single table) -> CSV Writer (b) ... Loop End
In order to have dynamic name I also have Java Edit Variable (Simple) from the Flow Variable port of (a) which follows s.roughley's suggestion in this thread http://tech.knime.org/forum/knime-users/writing-multiple-files
where I'm replacing the .txt read in with .csv
and i can see in the snippet output it has worked.
In the CSV writer node I can select 'Output File Locaion:' via the 'v=' button as Flow Variable, and I can select the filename flow variable I've created in the java snippet, however I get two issues;
If the actual text box is empty (even after selecting the variable) I'm unable to leave the configure menu with a "Invalid Settings: Missing Output file name error".
If put any text in the text box, and then execute, I get an "Execute failed xecute failed: Unable to create directory for specified output file: [some path I've not entered such as the workspace] [the path from the flow variable]". So it looks like the CSV node is concatenating the variable I've provided with [some path], rather than just using my flow varibles path.
So, something I'm doing I guess - how can I correctly reference the flow variable created by the java snippet in the csv writter node so that the it uses this as the filename?
Cheers,
Andy