A five-minute check before your next CSV import
Read the structure first. A clean-looking spreadsheet can still carry an extra separator, an empty header, or a blank required value.
Keep an untouched source
Work from a copy exported by the source system. Give it a descriptive filename and keep the original outside this workspace. A downloaded result should be a new version, not a substitute for your only source. Rowlane keeps input in the current page memory and does not save a history you can retrieve later.
Check the header and separator
The first CSV record becomes the header. If a table you expect to have five columns appears as one column, check the delimiter selector before doing anything else. A semicolon-separated export is not a malformed comma-separated export; it is a different dialect. Duplicate or empty headers need correction because a transformation cannot know which unnamed field you intended.
Read findings as questions
For id,name with 001,Avery and 002, on separate records, the last name is empty. That is a useful finding, but it only becomes an error if your destination requires a name. An exact repeated record is worth investigating; it may be a duplicated export or a legitimate repeated event. Record numbers include the header, and multiline quoted fields count as one record.
Resolve structure before changing values
An uneven record can come from an unquoted comma inside a name. Open the source export and compare that record with its header. Do not add or delete values simply to make the count look right. After correcting the source, run the check again, then move to cleaning or validation. Keep the findings report with your import notes.