ABOUT COLUMN HARBOR

Useful checks.
Explainable results.

We help people import, clean, compare and export everyday data with Excel, Google Sheets and small, explainable file workflows.

Who makes this

Column Harbor is an independent project by JHSOFT, a software business in South Korea. Its founder has 11 years of software development experience. This site is not affiliated with Microsoft or Google.

The people, other projects and Korean data guides behind JHSOFT are introduced on the JH Soft Labs homepage ↗. Column Harbor is its English CSV and spreadsheet learning site.

We focus on repeatable data problems: identifiers that lose zeros, dates that change meaning, hidden spaces, mismatched exports and file conversions that need clear rules. The checker is free and currently has no advertising or affiliate links.

How we produce a guide

Start with a concrete problem and an original synthetic dataset.
Check software behavior against official documentation and link to it.
Check parser behavior, sample calculations and runnable code with representative inputs and failure cases, where applicable.
State the limits, especially where an import choice depends on your data or Excel version.

Writing and software development are AI-assisted. We do not claim that AI output or automated tests replace hands-on testing in every target application. The guides distinguish our parser checks from behavior documented by the relevant software publisher. Data diagrams are illustrations of our examples, not screenshots of an application. Published and updated dates describe the article history, not continuous monitoring.

Examples you can reproduce

A plausible answer is not enough when a spreadsheet can silently change a value. These guides include the inputs, the expected result and the limits of our checks. Native tests below used Excel 16.0 build 5569.0; they do not establish behavior in every version.

Regional dates and amounts: four records, three recorded amounts and a total of 1208.75. We ran the query and tested invalid dates, invalid amounts and exact-format rejection on October 5, 2026. Download the full query and result checklist.
CSV files with changing headers: three files produce four rows totaling 13. Native tests on October 5, 2026 verified reordered columns, missing optional values and rejection of broken required headers. The article records an older-engine fix.
Monthly columns to rows: the query was executed in Excel on October 3, 2026, then refreshed with an additional month to check that new values were included. The guide provides the input CSV, M source and expected output.

Engine execution is different from clicking through the user interface. Each guide records that boundary. A correction changes the original article and its update date; we keep its URL so earlier links continue to reach the maintained version.

What the tool can tell you

It can show values present in a UTF-8 source file and flag selected patterns. It cannot restore overwritten values, validate every destination schema, scan every encoding or certify a file as safe. Formula-like prefixes are flagged as a precaution, but this is not a complete security scanner.

Corrections and feedback