A tool for choosing which columns to keep and changing their order.
Organizing CSV columns is simple in principle, but it takes time.
You change the order of columns or remove the ones you don’t need. You work through them one at a time in spreadsheet software, then check again at the end to see whether any values have changed. Even when you are used to it, the more you repeat it, the heavier it feels.
The tool introduced on this page handles just that column-organizing step. It loads a CSV in your browser, lets you select the columns to keep and change their order, and then saves it as a new CSV. It does not rewrite values.
Tool location
CSV Column Reorder & Select
https://wp159995.wpx.jp/tools/csv-columns/
It cannot filter rows, transform values, or rename columns. The limits are 20MB, 100,000 rows, and 200 columns.
What this tool can do
• Load a CSV in your browser by selecting a file or dropping it onto the tool. You can also try it with the sample data.
• Select the columns to keep.
• Change the order of the columns.
• Save the result as a new CSV.
• Before saving, the tool rereads the generated content and compares every cell with the original. If there is a mismatch, it does not save.
How to use it
1. Load a CSV
Select a file or drop it onto the tool. You can also try the sample data.
2. Choose the columns to keep and their order
Select the columns you want to keep and change their order.
3. Check the result and save
Before saving, the tool rereads the generated content and compares every cell with the original. If there is a mismatch, it does not save.
Limitations and size limits
• Row filtering is not supported.
• Value transformation is not supported.
• Column renaming is not supported.
• Only comma-separated files are supported. Tabs and semicolons are not supported.
• Limits: 20MB, 100,000 data rows, and 200 columns.
• The character encoding is UTF-8. If the file cannot be read, you can reread it using Shift_JIS.
What to know before using it
For a CSV with only one column, if the last row contains an empty value without quotation marks, that empty value is ignored when the file is read. This is because it cannot be distinguished from a trailing blank line.
Data processing and network activity
This tool is built to load and process CSV files in the browser.
Separately from its design, we also observed it in an actual browser after placing it on the site. During the check on November 6, 2026, only files from this site were loaded, with 0 external resources.
This is a record limited to what we could observe. We do not know what happens under conditions we have not checked.
Clear Conditions Board
What we have checked:
• The checking side (a separate AI from the one that built it) independently reran the builder’s 22 processing tests and the 9 tests after the fix. The results matched.
• The first acceptance test found a defect in which a row containing an empty value explicitly enclosed in quotation marks disappeared when read. We fixed it and ran acceptance testing again. It passed within the specified scope.
• On November 6, 2026, we used fictional sample data with 6 rows and 7 columns in an actual browser. We were able to remove columns and change their order. The comparison before saving covered 36 cells, with 0 mismatches.
• We were able to reread a Shift_JIS file by following the guidance.
• A file larger than 20MB displayed a message saying the file is too large (“大きすぎます”) and stopped. A malformed CSV also displayed a message indicating that processing had stopped.
• Only files from this site were loaded, with 0 external resources. This result is limited to what we could observe.
What we have not yet checked:
• Use with actual work data.
• Speed with large files.
• Behavior under conditions not included in these tests and observations.
Who is building it
I’m Rook, the person behind SAMUHARA. I’m building this tool with AI companions.
We separate the role of building from the role of checking. We will keep a record of the failed first acceptance test, as well as the fix and the checks that followed.
Current publication status
Checked on: November 6, 2026
The tool is placed on the site at its initial domain. This introduction page and the article documenting its creation (Quest Log) were published on November 6, 2026. The site is set to appear in search results.
This English version was translated with the help of AI.