Introducing Caxton 2.0. More powerful than ever.

1 of 7

A 7.7 GB CSV opens into the grid.

Open the file and the grid is there. The index of every row finishes in the background, and the status bar reports its progress while you read.

Grid on screen 1.6 s after Open.
Full index of 50,003,363 rows: 21.1 s.

2 of 7

Transform millions of values with confidence.

The new Transform Cells feature provides 48 transforms for cleaning and modifying data. Users can trim whitespace, change text case, replace or extract values using regular expressions, convert dates, round numbers, normalize boolean values, and more. Transforms can be applied to individual cells, selections, entire columns, or complete tables, and under a filter, to the visible rows only.

Before applying a transformation, the Transform Cells sheet previews the original and resulting values for a sample of the first fifty, along with how many of the sample values will change. After Apply…, a confirmation gives the full count of cells that will change, stay the same, or cannot be converted, and warns when a batch passes 10,000 cells. Values that cannot be converted remain unchanged and are reported rather than silently modified.

Even when a transformation affects millions of cells, the entire operation can be reversed with a single Undo.

Row filter across the file: 40.0 s.
Upper Case on 24,943 visible cells: under 2 s.

3 of 7

Find problems before they become bigger problems.

Caxton 2.0 introduces schema-based validation, allowing users to define exactly what their data should contain. Validation rules can specify column types, required values, uniqueness, numeric ranges, regular expression patterns, and lists of permitted values.

Structural checks cover eight rules: column count, quoting, escapes, line endings, encoding, duplicate headers, missing headers, and empty records. The run in this clip adds one column rule to the eight.

Validate all 50,003,363 rows: 255.6 s.

4 of 7

Go straight to each problem.

Every issue appears in the new Problems panel, where users can navigate directly to the affected record, inspect a suggested repair, and preview the correction before applying it. Double-click, Return or Go to Problem jumps to the row, and Back returns you to where you were.

Malformed records can also be quarantined into a new document, with three added columns (problem_rule, problem_message, and original_row) that preserve the reason for each failure.

In testing, Caxton identified 505,835 problems in one rule across 2,022,947 rows. The run in these clips found 1,494 problems in 4 rules across all 50,003,363 rows.

A repair across the whole file streamed in 82.8 s.

5 of 7

Understand your data at a glance.

The new Column Profile provides a detailed view of the contents of any column. Caxton calculates data-type distributions, blank values, distinct counts, numeric and date ranges, and value frequencies. Results can be searched, exported, or used to filter the table.

Column profiling is designed to remain efficient with large datasets, with frequency data stored in temporary disk space rather than retained entirely in memory.

A column of one million distinct values on a 1 GB file: about 15 s.

6 of 7

Remove duplicates, even across 50 million rows.

Table Mode also gains advanced duplicate removal and external sorting for datasets that exceed available memory.

Users can identify duplicate groups and choose which records to retain, including the first or last occurrence, the earliest or latest by date, or the highest or lowest numeric value, or keep none of them. Caxton reports duplicate groups before removing records and provides the option to export the results into a new document.

Scan on three columns: 73.9 s.
2,867 duplicate groups.

7 of 7

The same file as plain text.

One switch shows the CSV as text, with Caxton's text tools: find, filter, bookmarks and multiple cursors. Switch back and the grid returns.