Skip to content

Table Tools

A few quick utilities for reshaping and tidying the active table.

Transpose

Analyse > Transpose swaps rows and columns into a new tab. The original column names become the first column, and each original row becomes a column. Everything is shown as text, since a transposed mix of column types has no single type.

Transpose is limited to tables of at most 1000 rows, because each row becomes a column and a very wide result would be unusable. Above that limit the status bar tells you instead of running.

Compare rows

Analyse > Compare rows... puts the rows you picked side by side so you can see where two records, or ten, actually disagree.

Picking them takes either gesture:

  • Select the rows (click a row number, then Ctrl+click or Shift+click the others). This is usually the quickest way in.
  • Mark them (right-click a row number > Mark, or Ctrl+M), which survives scrolling away and clicking elsewhere.

The selection wins whenever it holds two or more rows, so a set of marks you built up on purpose is not quietly replaced by a stray click; with fewer than two rows selected, the marks are used. The entry stays greyed out until either gesture has picked at least two rows, because one row differs from nothing.

The result opens in its own tab, one output row per column of the original:

column differs row 12 row 4711
city no Aachen Aachen
amount yes 12.50 13.10

Run your eye down differs to find the fields that are not the same in every picked row. Every field answers, yes or no, so a blank cell never has to be read as "not checked". Values are compared exactly as they are displayed, so a flagged field always looks different on screen too. Marked cells and marked columns are ignored: only whole-row marks pick a row for the comparison.

Random sample

Analyse > Random sample... opens a small dialog where you type a row count (100 by default). It then opens a new tab holding that many rows, chosen at random from the active table. This is handy for eyeballing a fair cross-section of a big file without scrolling all of it. If you ask for more rows than the table has, you get them all.

On a database tab that holds only part of its table, the database picks the rows: Exact or Fast, see Analyses on live databases.

Tidy up

Data > Tidy up... cleans the current table in a single undoable step. Two options:

  • Trim spaces from cells and headers (on by default): removes leading and trailing spaces from text cells and column titles. Spaces inside a value are left alone.
  • Tidy column names to snake_case (off by default): lowercases column names and replaces runs of punctuation or spaces with a single underscore.

A single Undo reverts the whole thing. Octa can already do this automatically when opening a file; this lets you run it at any time.

Command line and MCP

The same job without the GUI: