Start with the task at hand

Start with a concrete output: one combined table, a list of differences, a smaller file, or a file ready for another application. Two tasks can use the same input format while needing very different operations. Appending monthly exports increases the number of records. Matching a product list to a price table adds information to existing records. Comparing yesterday’s export with today’s identifies changes. Choose the operation that answers your question before choosing files. The directory groups related tasks and lets you search by tool name, description, or a familiar term in English or Simplified Chinese.

Most tools begin with a file picker or pasted text, but supported formats vary. The format labels on each card describe the input for that particular tool. Dedicated workbook tools handle XLSX files, while the nested-data converter accepts JSON and JSONL. Calendar and contact tools turn supported table inputs into another format. A CSV-only page does not gain workbook support because another page has it. Open the matching tool and check its input options before preparing a large file.

Choose the cleaner for selected text changes such as trimming surrounding whitespace, adjusting letter case, or handling an explicit missing marker. Keep identifiers outside a cleaning rule unless the transformation is intentional. An account code can contain meaningful spaces, and an empty cell is different from a zero. Preview the changed-cell report to see which values the rule affects. Cleaning is a useful first step when superficial differences prevent records from matching, but it should not erase distinctions that matter to your work.

Use duplicate removal when you need to keep or mark records with the same selected key. Use duplicate-record consolidation when several records describe the same entity but contain complementary fields. These tools answer different questions: one chooses surviving rows, while the other assembles fields and exposes conflicts. A shared name alone may not identify the same person or company. Fuzzy matching offers candidates for review when text is similar; it does not establish identity or make those decisions for you.

Merge appends rows from selected files, aligning fields by confirmed column names. It can preserve all fields, require the same fields, or keep only shared fields after confirmation. Join matches keys to bring fields from another table into the result. A repeated key can produce several matches and more output rows than expected. Inspect duplicate-key and unmatched-record reports before accepting the relationship. Neither operation updates a database or changes a source file on your device.

Compare finds added, removed, modified, and unchanged records using the selected matching rules. A stable key lets records match even when their positions change. Workbook comparison also requires you to pair sheets and decide whether to compare stored values, displayed text, or formula text. Three-way merging serves a different situation: two people edited separate copies of the same original. It uses that original to distinguish independent changes from conflicts that need a choice.

Splitting a column creates several fields from one value. Splitting a cell into rows creates several records and repeats the other fields beside each new item. That repeated context is useful for survey answers or item lists, but summing a repeated amount afterward can overstate a total. Concatenation combines selected fields into text in an explicit order. Extraction takes a substring using the supported character or literal-marker rules. Keep the source column when you still need to compare the result with the original text.

Unpivot turns selected wide columns into attribute and value records while keeping identifier columns. Transpose swaps row and column positions. Pivot summarizes values into a cross-tabulation. These operations have different effects on record count, aggregation, and header meaning. The relevant page explains those effects and provides examples. Multi-row header flattening is useful when a workbook begins with layered headings that must become a single set of field names before further processing.

Date and number conversion need an explicit interpretation. The text 03/04/2026 can describe different dates, and a comma can be a decimal separator or a grouping separator. Interface language does not decide the meaning of your data. Confirm the source convention, inspect unresolved values, and choose the output format. Category standardization works from mappings you approve, allowing variant spellings to be grouped without silently translating unfamiliar values or treating similar names as equivalent.

Use profiling to understand missing values, distinct values, frequencies, and confirmed numeric statistics. Validation checks chosen rules such as required fields, uniqueness, ranges, and relationships between fields. Grouping and pivoting then summarize the records under explicit aggregation choices. A passed check means the tested rules passed on the supplied data. It does not prove that a source export is complete or that an external application will accept every row.

The column calculator produces new values or conditional results from configured operations. It does not execute arbitrary spreadsheet formulas pasted into a field. Running calculations add cumulative totals, previous-row differences, or rolling values while preserving detail. Their meaning depends on grouping and order. A window of seven records is different from seven calendar days when dates are missing or multiple records share a date.

The missing-date tool examines an intended grid and can create missing records after you review its size. New records do not automatically imply zero sales or zero measurements. Historical matching finds the record applicable at a particular time, using the configured date rule. Reconciliation helps review relationships between transactions and settlements. Check the exceptions and confirmation state rather than treating every suggested relationship as an accepted match.

For specialised outputs, choose the calendar converter for ICS events, the contact converter for VCF contacts, or the Shopify validator for supported product CSV checks. Each destination has its own conventions, and generating a file does not send it to an account. Read the format-specific limits and test a small import in the destination before using the complete result. Conversion cannot guarantee how every calendar, contact manager, or store configuration will interpret the file.

All tools process file contents in the browser. Save results you need before refreshing or closing the page. Shared table workflows can carry an applied result between compatible tools; specialised workspaces may use separate sessions or require an exported handoff. Follow the options shown on that page. Downloads create new files, and a result table, exception report, and rule file serve different purposes. Keep the files needed to understand and reproduce your chosen transformation.