Convert TSV to PostgreSQL SQL

How to convert a TSV file to PostgreSQL

A TSV file is just like a CSV, except columns are separated by a Tab character instead of a comma. SQLified detects that automatically — drop the file and it reads each tab-separated column, infers a PostgreSQL type, and writes a clean CREATE TABLE + INSERT script. No need to re-save your data as CSV first, and no database access required to build the SQL.

1. Drop your TSV onto the converter above (.tsv, .tab, or tab-delimited .txt all work).
2. Review the schema. Adjust any inferred type, mark columns nullable, rename them, or set a primary key.
3. Generate the SQL — a CREATE TABLE plus batched INSERT statements.
4. Run it in Postgres — paste into psql, pgAdmin, DBeaver, or your Supabase / Neon / RDS console.

Where TSV files come from

Tab-delimited data shows up everywhere once you look: Excel's “Save as → Text (Tab delimited)”, Google Sheets exports, mysqldump --tab, PostgreSQL's own \copy ... TO, and a huge amount of scientific and bioinformatics data (.tsv / .tab). Because tabs almost never appear inside real values, TSV sidesteps the comma-escaping and quoting problems that make messy CSVs painful — which is exactly why so many exports use it.

What the generated SQL looks like

A tab-separated file with columns sample_id, label, collected_on, reading becomes:

CREATE TABLE IF NOT EXISTS "samples" (
  "sample_id" INTEGER NOT NULL,
  "label" TEXT NOT NULL,
  "collected_on" DATE NOT NULL,
  "reading" NUMERIC(18,3) NOT NULL
);

INSERT INTO "samples" ("sample_id", "label", "collected_on", "reading") VALUES
('1', 'control', '2026-01-04', '0.482'),
('2', 'treated', '2026-01-04', '1.139');

Identifiers are double-quoted (the PostgreSQL standard) and CREATE TABLE IF NOT EXISTS lets you re-run the script safely.

PostgreSQL type inference

SQLified reads the actual values in each tab-separated column and maps them to the right Postgres type:

• Whole numbers → INTEGER, or BIGINT beyond the 32-bit range.
• Decimals → NUMERIC(18, n) sized to your data's precision.
• Dates and timestamps → DATE or TIMESTAMP.
• true/false → BOOLEAN.
• Everything else → TEXT.

Override any guess before generating — handy for sample IDs or codes you'd rather keep as text.

COPY with a tab delimiter vs. INSERT

PostgreSQL can load a TSV directly with COPY samples FROM 'file.tsv' WITH (FORMAT csv, DELIMITER E'\t') — but that needs the file to be reachable from the database host, which managed services like Supabase, Neon, and RDS make difficult.

SQLified instead produces portable INSERT statements you can paste into any SQL console. For large exports it splits them into 1,000-row batches so they import reliably rather than failing as one oversized statement — the simplest path when you only have a SQL editor, not file access.

Frequently asked questions