Short answer
What you need to know
Read the CSV as rows and columns, search only inside the values, and write each changed value back with the quoting the format needs. Then commas between columns and quotes around values stay in place, and removing every comma or quote from the values still leaves a valid file with the same rows and columns.
Why plain find and replace breaks a CSV
In a CSV, a comma separates columns, and a value that itself contains a comma is wrapped in double quotes, like "Smith, John". A double quote inside a quoted value is written twice.
Replacing every comma in the file joins columns together. Removing every double quote un-wraps "Smith, John", and the comma in the middle then splits it into two columns. Either way, rows end up with the wrong number of columns and the import fails or, worse, stores values in the wrong columns.
Change the values, not the file
Data Check first reads the file the way a database does, so it knows where each value starts and ends. The characters and phrases you enter are looked for only inside values.
When a value changes, it is written back with quotes if it needs them. Values that did not change, and the separators between them, are copied exactly as they were.
Type the characters you do not want together, such as " ; , -, and each is looked for on its own. Words and phrases go one per line, and can be matched only when they are the whole value, which suits placeholders such as NULL or N/A that should become empty.
Allow only what a column should contain
Sometimes it is easier to say what is allowed. A phone column should hold digits; a code column, letters and digits. Turn on Only allow, pick the groups and any extra characters, choose the column, and every other character is reported and can be removed.
Check the list of characters found before removing them: a + in a phone number or an accented letter in a name may be correct.
A worked example
A notes column holds "He said ""hi""; then left - fast". Entering " ; - in Data Check reports two double quotes, one semicolon, and one dash in that column.
Choosing Remove for each, with Double spaces ticked, gives He said hi then left fast: the space left behind by the dash is tidied up last. The file still has the same columns in every row, and the cleaned value keeps the quotes it was written with.
Important limitation
Verify important findings
Removing characters changes the data. Keep the original, review the before-and-after list, and limit the cleaning to the columns that need it, since a character that is wrong in one column can be correct in another.
Remove characters from a CSV with Data Check Review the methodology
Further reading
Authoritative references
These primary sources support the concepts and reference information in this guide. Their inclusion does not imply endorsement of Hidden Word Scanner.