Keep a partial header-like data row

Partial matches stay in the table

Sometimes a real record uses a word that also appears in the header. A product could be called Name, a customer could be listed as Total, a category could be Amount. A cleanup that removes every row containing a column name would delete those records along with the repeated headers, and the loss would only show up later as a missing line in a total.

Repeated header cleanup avoids that by comparing whole rows. A later row is removed only when every one of its cells equals the corresponding cell of the chosen header, after trimming outer spaces.

Example: Name with 12.50

The page example has three source rows: Name and Amount, Name and 12.50, and Name and Amount again. With Header row set to 1, the cleaner removes only source row 3. Row 2 is kept because its second cell, 12.50, differs from Amount. Removed header rows shows 1 and the output has 2 rows: the header and the record with 12.50.

The kept row is unchanged. The cleaner does not flag it as suspicious or modify its first cell.

Had the record read Name and Amount in both cells, it would have been removed along with the repeated header, because identical text gives the cleaner no way to tell intent apart.

Reviewing what was kept

• After Calculate, compare Removed header rows with the number of repeated headers you expected.
• Find records whose first cell matches a column name and confirm they carry real values in the other cells.
• Check that the removed row really was a full header, not a record whose values happen to equal the column names.
• If a row looks like a broken header, such as Name with an empty second cell, decide from the source whether it is a header fragment or a record with a missing amount.
• Keep your original text so you can repeat the comparison.

Why a keyword search is the wrong tool

Searching for the word Name and deleting matching rows feels faster, but it cannot tell a header from a record. If you need to remove something the cleaner kept, such as a header with a page number added to one cell, edit that row in a copy so it matches exactly, then calculate again.

In the Tablora app, Merge pages applies the same whole-row principle when it drops a repeated header at the top of a later page: every cell must match, ignoring letter case and outer spaces. A record that shares only one cell with the header is appended like any other row.

Original synthetic example

Browser header row

1

Removed source rows

3

Output rows

2

Tools

• repeated-header-cleanup