I am getting NullPointer Exception Error after about 19% processing by the Term co-occurrence counter.
The workflow is as follows:
XLS reader -> Strings to Document -> Simple Processing -> POS Filter -> BoW creator -> Term Co-Occurrence Counter
Manual verification of the terms column indicates only presence of hyphenated words. I was wondering as to why this error could arise.
Also, is there a dehyphenator node? I found a hyphenator node. I was wondering whether there is a node with opposite function.
Thanks in advance.
By trial and error, I found that the term cooccurrence node takes up to 50,000 rows. I am not sure whether it is an in-built limitation or the limitation of the configuration of my system [laptop].
Please clarify as and when you could.
which version of KNIME are you currently using? There was an issue with NullPointerExceptions in 2.9.4 and earlier versions, which is fixed in 2.10 and later. When you update to 2.10.1 you need to re-execute the workflow after the XLS Reader node.
Alternatively you could set the File Store Chunk Size in the Preferences
to a number larger than your number of documents, i.e. rows of your XLS Reader. You need to re-execute the workflow after the XLS Reader node as well.
If you still get the Exception please post the StackTrace from the console. That would help a lot to identify the problem.