About

Data rarely arrives in a clean, database-ready format. Analysts and developers frequently encounter the Unstructured Data Problem: valuable information trapped in PDF tables, log files, or legacy document formats where the relationship between data points is defined visually rather than syntactically. This tool bridges that gap using heuristic parsing algorithms.

We employ a Matrix Transformation logic, denoted as f: S → M_m×n, where an unstructured string S is decomposed into a structured grid M. Unlike simple splitters, this engine respects RFC 4180 standards for quoted fields and employs Whitespace clustering to detect columns in visual-only formats (like PDF dumps).

Formulas

The core parsing logic differentiates between Delimited Parsing and Fixed-Width Heuristics. For delimited text, we define the splitting function:

{

Cell ← buffer + char if in_quotesNewCol ← PUSH if char ≡ delim ∧ ¬in_quotes

For PDF/Visual alignment, we utilize a Density Function ρ(x) where x represents the character index. Peaks in ρ(x) across multiple lines L_i indicate column boundaries:

Boundary ⇔ N∑i=1 is_space(L_i[x]) > Threshold

Reference Data

Format	MIME Type	Structure S	Use Case
CSV	text/csv	Row ⋅ DELIM ⋅ Col	Excel, Pandas, Legacy Imports
JSON	application/json	[{k:v}, ...]	Web APIs, NoSQL Databases
SQL	application/sql	INSERT INTO table	Relational DB Migrations
XML	application/xml	<root><row>...	Enterprise SOAP Services
TSV	text/tab-separated-values	\t Delimited	Clipboard, Unix Tools

Frequently Asked Questions

PDF text often lacks delimiters, relying on visual whitespace. Our "Multi-Space" delimiter mode treats consecutive spaces (regex /\s{2,}/) as a column break, effectively reconstructing the visual table structure into a logical grid.

No. This tool is RFC 4180 compliant. If a cell contains the delimiter or newlines, it will be automatically enclosed in double quotes. Existing double quotes are escaped as pairs ("").

Yes. The "Data Workbench" allows you to Transpose (swap rows/columns), delete specific empty rows, and reorder columns via drag-and-drop mechanics before generating the final output.

The generator scans each column. If all values are numeric, it assigns "DECIMAL" or "INT". If dates are detected (ISO 8601), it assigns "DATETIME". Otherwise, it defaults to "VARCHAR(255)".