Data tool usage reference
Choose the operation that matches your task. Merging appends rows; joining adds lookup columns; comparison reports changes. They solve different problems.
Image workflow
Batch image resizing, formats, padding and ZIP download →
Before you begin
Keep the original file. Use UTF-8 where possible, and choose the actual CSV delimiter. Standard input limits are 5MB, 20,000 data rows and 200 columns; XLSX input is 2MB. Selected files are processed in browser memory. CSV export includes a BOM, and formula mitigation can change values beginning with =, +, -, @ or whitespace followed by those characters.
Excel to CSV
Extract saved values from an XLSX worksheet and download UTF-8 CSV.
Select a worksheet. Only stored values and cached formula results are read; formatting, macros, encrypted files and legacy XLS are unsupported. Dates and percentages may appear as raw values. XLSX input is limited to 2MB and 20MB uncompressed.
CSV to JSON
Convert CSV rows into a JSON array while preserving identifiers as strings.
The first row supplies object keys. Empty or duplicate headers and uneven rows are rejected. Every value, including numbers and blanks, remains a string; blanks do not become null.
CSV to SQL INSERT
Generate reviewable INSERT statements for PostgreSQL, MySQL or SQLite.
Use a table name with letters, digits and underscores. Values are quoted strings, including empty strings. Types and constraints are not inferred. MySQL output sets NO_BACKSLASH_ESCAPES; PostgreSQL values containing backslashes are rejected. Review the generated SQL in a test database. Nothing is executed here.
Remove duplicate CSV rows
Keep the first or last complete row for each key, in original source order.
Key columns are 1-based: use 1 or 1,3, or leave blank to compare complete rows. Empty keys also match. Case and leading zeros matter. Ignoring surrounding whitespace affects matching only. Last means file order, not the latest date; values from discarded rows are not combined.
Merge CSV files
Append CSV files with matching headers into one table.
Select 2–20 files, with a combined size up to 5MB. Encoding, delimiter, header names and column order must match. Rows are appended in the displayed file order; the header appears once. This operation does not join by key or remove duplicates.
SQL IN list generator
Turn one value per line into a quoted SQL IN list.
Blank lines and exact duplicates are removed. String mode preserves codes such as 0012 and doubles single quotes. Numeric mode accepts integers and decimals only. Check database list limits and MySQL backslash settings before running SQL.
Compare two CSV files
Find added, removed and changed records using a unique key column.
Both tables need identical headers and column order. Key values must be nonempty and unique in each table. A changed row can produce several changed-cell records. Whitespace comparison is optional. Reports are limited to 20,000 rows and 20MB.
CSV lookup and JOIN
Add matching lookup columns using a LEFT or INNER join.
Choose the 1-based key column in each table. Lookup keys in table B must be unique and nonempty. LEFT keeps unmatched A rows with blank lookup values; INNER drops them. This is an exact lookup, with optional surrounding-whitespace matching.
Edit CSV columns
Select, reorder and rename columns without changing cell values.
Load the column list, select columns to include, edit their names and move them with the arrow buttons. At least one column must remain. Output headers must be nonempty and unique.
JSON to CSV
Convert a flat array of JSON objects into a rectangular CSV table.
Input must contain 1–20,000 objects. The union of object keys becomes the header. Missing keys and null become blank cells. Nested objects and arrays are rejected; quote large integer identifiers to preserve them.
CSV to Excel
Create an XLSX workbook with text cells to preserve identifier values.
The CSV needs nonempty, unique headers and consistent row widths. Cells are stored as text rather than inferred numbers or dates. Formatting and formulas are not recreated; formula-like text may receive a protective apostrophe.
JSON formatter and validator
Validate JSON syntax, pretty-print it or remove unnecessary whitespace.
Use standard JSON with double-quoted keys and no trailing commas. Unsafe integers must be quoted as strings. Formatting does not sort object keys or validate a business schema. Output is limited to 20MB.
Transpose CSV
Swap rows and columns in a small table.
The entire input, including the header, must have at most 200 rows because transposed output is limited to 200 columns. This swaps cell positions without inferring a new header or converting types.
Split CSV into ZIP
Split data rows into smaller CSV files with a repeated header.
Enter 1–20,000 data rows per part. Each file includes the original header. Output is limited to 100 files and 20MB, including repeated headers. The preview shows the first part; download contains all parts as ZIP.
Remove duplicate text lines
Keep the first occurrence of each nonempty line.
Empty lines are removed and comparison is case-sensitive. Optional trim affects matching only; the original first line is kept. Up to 20,000 lines and 5MB are supported.
Clean CSV whitespace
Trim surrounding cell whitespace and remove empty data rows.
Trimming includes header cells. Blank-row removal deletes rows whose cells are all empty or whitespace. Nonempty malformed rows are rejected rather than repaired. Formula mitigation can add an apostrophe; disable it only after reviewing the values and destination.
CSV and TSV converter
Convert comma-separated and tab-separated tables with proper quoting.
Select the actual input delimiter, then the target format. Quoted fields can contain delimiters and line breaks. CSV and TSV downloads include a UTF-8 BOM. Cell values remain strings.
CSV to Markdown table
Build a Markdown table with escaped cell contents.
The first row becomes the header. HTML is escaped, pipes and Markdown punctuation are protected, and cell line breaks become br tags. Editors that disable inline HTML may display those tags literally.
Sort CSV rows
Sort one column as text or exact decimal values while keeping ties stable.
Column numbers start at 1. Empty or whitespace-only values remain last in either direction. Text uses Korean locale collation, including for English input. Numeric sorting accepts plain integers and decimals without units or grouping commas. Dates are not parsed as a dedicated type.
Filter CSV rows
Extract rows using one column and an exact or text-matching condition.
Choose contains, equals, not equals, starts, ends, empty or not empty. Optional case and whitespace normalization affect comparison only. No regex or numeric magnitude comparison is applied. Zero matches produce a header-only result.
JSONL and JSON converter
Convert one JSON value per line to an array, or an array to JSONL.
JSONL input must be UTF-8 without BOM. Blank lines are errors unless explicitly skipped; a final newline is allowed. The reverse direction requires a top-level array. Limits: 20,000 records, depth 128, 100,000 values and 20MB output.
Compare JSON structures
Find added, removed and changed values using JSON Pointer paths.
Object key order and indentation are ignored. Arrays are compared by position. Missing keys, null and different types remain distinct. An empty path means the entire document. Added or removed objects are reported as one change. Limits: 5,000 changes, depth 128 and 20MB output.
Review the saved file
Check headers, row counts, blank cells and identifiers such as 0012. Spreadsheet software may automatically interpret text as numbers, dates or formulas. Import code columns explicitly as text. Results do not preserve the original spreadsheet formatting.