Hi, I also faced similar issues while applying regex_replace() to only strings columns of a dataframe. The trick is to make regEx pattern (in my case "pattern") that resolves inside the double quotes and also apply escape characters. val pattern="\"\\\\[\\\\x{FFFD}\\\\x{0}\\\\x{1F}\\\\x{81}\\\\x{8D}\\\\x{8F}\\\\x{90}\\\\x{9D}\\\\x{A0}\\\\x{0380}\\\\]\"" val regExpr = parq_out_df_temp.schema.fields.map(x => if(x.dataType == StringType){"regexp_replace(%s, $pattern,'') as %s".format(x.name,x.name)} else x.name) val parq_out=parq_out_df_temp.selectExpr(regExpr:_*)   This worked fine for me !! ... View more