DataSet change in Quicksight

Hi, I have created a dashboard using dataset-1 which has 20 columns and later new data has 22 columns in it. The datasets in QuickSight are created using s3 data. Now whenever I refreshed the dataset, the rows are getting skipped because of incorrect number of rows/data-types. How should I resolve this problem?
Thanks in advance!

If you are using athena or redshift spectrum, you can just re-save the dataset and it will be fine. If you directly load the data from s3, you have to create a new dataset and then replace the old dataset in analysis. You can remap the columns.

Thanks for your response! It meant a lot for me.
But our file format in s3 is [{“col1”:“val1”, “col2”:“val2”…},{},{}…]. I am not able to collect that data from s3 to athena properly as the data format is not a proper json.
Could you please help me on this. Thanks in advance!