Skip to content

Header row

Some files do not start with their column names. Bank and government exports often put a title block above the real header, Excel reports spread the header over two rows with merged cells, and some files have no header at all. Octa opens them as they are and lets you fix the header afterwards.

Use a row as the header

Right-click any cell in the row that holds the column names and pick Use as header row. That row becomes the column names, and it and every row above it are removed.

When the header spans two rows (a merged 2025 above Q1 Q2 Q3 Q4), select both rows first: click the first row number, Shift+click the second, then right-click. The entry reads Use rows 4-5 as header and combines them. A blank cell in the upper row takes the text to its left, so the merged 2025 reaches every quarter below it and the names become 2025 Q1, 2025 Q2, and so on.

Blank names become column_1, column_2, ...; a name that repeats gets _2, _3.

Afterwards Octa looks at every column again, the way it does when a file opens: a column that was text only because of the title rows becomes numbers or dates again, European numbers like 1.234,56 included. A column that could be read two ways, such as dates that work day first and month first, stays text; Change column type lets you pick.

Ctrl+Z puts everything back in one step: the rows, the names and the types.

Octa suggests the row

When most column names are placeholders such as column_3 or Column3, the file almost certainly has a title block. Octa then looks through the first 20 rows for one that is filled across and made of words, and shows a banner:

The column names look like they are in row 4. Use row 4 as header Dismiss

Nothing changes until you click. For a merged two-row header the banner offers both rows.

A file without a header

If the first row of the file is data, Octa has used it as the column names. Untick Edit > First row is header: the names go back into the table as the first row, and the columns are called column_1, column_2, and so on. Tick it again, or press Ctrl+Z, to undo that.

When it is not available

  • In read-only mode and in large-file mode, like every other edit.
  • On database tables: their column names come from the database.
  • While a filter hides rows: rows above the header that you cannot see would be removed. Clear the filter first. A sort is fine, because the rows above are then the rows you see above.

Next month's file

Both actions are recorded as Recipe steps (use_header_rows, header_is_data), so replaying the recipe on next month's export fixes its header the same way: in Octa, with octa --recipe, and over MCP. The step remembers the row number, so it fits files whose title block has the same height.

Title lines in CSV files

A CSV whose first line is a title (Report generated 03.10.2026) used to open one column wide, because every row was sized to that line. Octa now sizes the table to the rows below it and finds the separator from the lines that agree with each other, so a ; file with a title line opens with every field in place. Columns that are empty in every row, because each line ends in a separator, are left out as before.