Hello @Jowawa and welcome to the KNIME Community
I would recommend to include some regex code in your filtering, in this way you can asume more possible string issues assumptions. This is: position in the string sentence, case sensitive, wildcard word combinations…
Please have a look to the following workflow example for insights:
Five starting nodes from this workflow are enough to solve your challenge description; adding a filter on ‘Trues’
Please comment if further support is needed, aiming to adapt it for your use case.
BR