Filtering XSUM dataset for a specific word in a column

Greetings everyone, I’d like to ask whether there is a way to choose a row from specific columns with some certain condition.

I’m using the XSUM dataset for summarization and I’d like to filter the dataset (the train split to be precise) where the rows that contains “environment” word is selected and the rest is discarded.

I’ve searched several methods such as converting the dataset into pandas dataframe format or even using the processing methods provided by the HuggingFace (Process) but I still unable to do this.

With that said, I kindly ask for any solution that can help me or even some insight is highly appreciated. Thank you everyone.