---
title: "INSERT"
slug: "dev-database-reference-sql-insert"
updated: 2026-10-30T18:19:11Z
published: 2026-10-30T18:19:11Z
canonical: "docs.vastdata.com/dev-database-reference-sql-insert"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vastdata.com/llms.txt
> Use this file to discover all available pages before exploring further.

# INSERT

*Applies to VAST Cluster 5.5.*

Inserts rows into a table, either as values or as the result of a `SELECT`. To import Parquet files with `INSERT`, see [import Parquet files](/develop/docs/dev-database-reference-sql-import-parquet).

## Syntax

```
INSERT INTO "<bucket>/<schema>".<table> VALUES (<value>, ...) [, (<value>, ...) ...]
INSERT INTO "<bucket>/<schema>".<table> (<column>, ...) VALUES (<value>, ...) [, (<value>, ...) ...]
INSERT INTO "<bucket>/<schema>".<table> [(<column>, ...)] SELECT ...
```

## Parameters and rules

| Part | Meaning |
| --- | --- |
| `VALUES (...)` | One row per parenthesised list. Without a column list, give a value for every column, in table order. Several rows in one statement work. |
| `(&lt;column&gt;, ...)` | The columns to fill, in the order of the values. |
| `SELECT ...` | Inserts the rows the query returns. The query must itself be one the Query Engine supports: one table, no `GROUP BY` ([SELECT](/develop/docs/dev-database-reference-sql-select)). |

- The statement returns one row with one column, `Count`: the number of rows inserted.
- Write literals as you would in a `SELECT`: `&apos;text&apos;`, `TIMESTAMP &apos;2026-01-01 10:00:00&apos;`, a list `[1, 2]`, a struct `{&apos;lat&apos;: 38.7, &apos;lon&apos;: -9.1}`. See [data types in SQL](/develop/docs/dev-database-reference-sql-data-types).
- The whole statement must fit the 5 MB statement size limit ([limits](/develop/docs/dev-database-reference-limits#sizes)). To load many rows, use ADBC bulk ingest (`cur.adbc_ingest()`) instead of long `VALUES` lists; the [profile a query](/develop/docs/dev-database-guides-profile-a-query#set-up) guide loads 200,000 rows that way. Bound parameters (`cur.execute(sql, params)`) aren't supported ([SQL reference](/develop/docs/dev-database-reference-sql-index#how-sql-reaches-the-engine)).
- There's no upsert: `INSERT OR REPLACE`, `INSERT OR IGNORE` and `ON CONFLICT` fail, because a table has no primary key. To replace rows, `DELETE` them and `INSERT` the new ones in one transaction.
- You can't insert into a table in a snapshot, or into a partitioned table ([not supported](/develop/docs/dev-database-reference-sql-not-supported)).
- On a sorted table, inserted rows are sorted later, in the background ([write path](/develop/docs/dev-database-concepts-architecture-write-path#sorted-tables)).

## What happens on the cluster

The Query Engine evaluates the values, or runs the `SELECT`, and writes the rows into the table on the cluster. With `INSERT ... SELECT`, the rows never leave the cluster: only the count comes back to your client ([ingest](/develop/docs/dev-database-guides-ingest#vast-query-engine)). For what happens to the rows after the commit, see [write path](/develop/docs/dev-database-concepts-architecture-write-path#inserts).

## Example

The examples use the `run()` helper and the environment variables described in the [SQL reference](/develop/docs/dev-database-reference-sql-index#example-connect-and-use-a-transaction). This block connects, and creates the schema `docs_sql_insert` with two tables:

```
import os
import re

import adbc_driver_manager
import adbc_driver_manager.dbapi as dbapi
import adbc_driver_vastdb

BUCKET = os.environ["VASTDB_BUCKET"]
SCHEMA = f'"{BUCKET}/docs_sql_insert"'

def connect(autocommit=True):
    return dbapi.connect(
        driver=adbc_driver_vastdb.get_driver_path(),
        db_kwargs={
            "vast.db.endpoint": os.environ["QE_ENDPOINT"],
            "vast.db.access_key": os.environ["AWS_ACCESS_KEY_ID"],
            "vast.db.secret_key": os.environ["AWS_SECRET_ACCESS_KEY"],
        },
        autocommit=autocommit,
    )

def server_message(e):
    """The server's message in a driver error, without the request ID or the echoed statement."""
    text = re.sub(r"\[req=0x[0-9a-f]+\]: ", "", str(e))
    found = (re.search(r'message: "(.*?)", request_id', text, re.S)
             or re.search(r"description: (.*)\)", text, re.S))
    message = found.group(1) if found else text.strip().splitlines()[-1]
    message = re.sub(r"^Plan query: '.*?' failed with: ", "", message, flags=re.S)
    message = re.split(r"\\n|\n", message)[0].replace('\\"', '"')
    return message.replace(BUCKET, "<bucket>")

def run(sql):
    """Run one statement. Print the rows it returns, or the error."""
    try:
        with conn.cursor() as cur:
            cur.execute(sql)
            rows = cur.fetchall() if cur.description else []
    except (adbc_driver_manager.Error, OSError) as e:
        print(f"{type(e).__name__}: {server_message(e)}")
    else:
        print(rows)

conn = connect()
run(f"CREATE SCHEMA IF NOT EXISTS {SCHEMA}")
run(f"CREATE TABLE IF NOT EXISTS {SCHEMA}.readings "
    "(id BIGINT, city VARCHAR, temperature DOUBLE, tags INTEGER[], location STRUCT(lat DOUBLE, lon DOUBLE))")
run(f"CREATE TABLE IF NOT EXISTS {SCHEMA}.warm (id BIGINT, city VARCHAR)")
```

```
[]
[]
[]
```

Insert two rows with all columns, then one row with a column list:

```
run(f"""INSERT INTO {SCHEMA}.readings VALUES
    (1, 'Lisbon', 21.5, [1, 2], {{'lat': 38.7, 'lon': -9.1}}),
    (2, 'Oslo', 4.0, [3], {{'lat': 59.9, 'lon': 10.7}})""")
run(f"INSERT INTO {SCHEMA}.readings (id, city, temperature) VALUES (3, 'Cairo', 30.2)")
run(f"SELECT id, city, temperature, tags, location FROM {SCHEMA}.readings WHERE id < 3 ORDER BY id LIMIT 10")
```

```
[(2,)]
[(1,)]
[(1, 'Lisbon', 21.5, [1, 2], {'lat': 38.7, 'lon': -9.1}), (2, 'Oslo', 4.0, [3], {'lat': 59.9, 'lon': 10.7})]
```

Copy the result of a query into another table:

```
run(f"INSERT INTO {SCHEMA}.warm SELECT id, city FROM {SCHEMA}.readings WHERE temperature > 20")
run(f"SELECT id, city FROM {SCHEMA}.warm ORDER BY id LIMIT 10")
```

```
[(2,)]
[(1, 'Lisbon'), (3, 'Cairo')]
```

An upsert fails:

```
run(f"INSERT OR REPLACE INTO {SCHEMA}.warm VALUES (1, 'Lisboa')")
```

```
InternalError: Binder Error: There are no UNIQUE/PRIMARY KEY Indexes that refer to this table, ON CONFLICT is a no-op
```

Clean up:

```
run(f"DROP TABLE IF EXISTS {SCHEMA}.readings")
run(f"DROP TABLE IF EXISTS {SCHEMA}.warm")
run(f"DROP SCHEMA IF EXISTS {SCHEMA}")
conn.close()
```

```
[]
[]
[]
```

## Errors

| You wrote | Message |
| --- | --- |
| `INSERT OR REPLACE`, `INSERT OR IGNORE`, `... ON CONFLICT DO NOTHING` | `InternalError: Binder Error: There are no UNIQUE/PRIMARY KEY Indexes that refer to this table, ON CONFLICT is a no-op` |
| An insert into a snapshot path | `InternalError: internal error: Access Denied` (status 1) |
| `cur.execute(sql, params)` | `ProgrammingError: INVALID_STATE: Cannot configure ingestion options on a statement configured for query` |

For a vector-indexed table, see the insert rules in [vector search SQL](/develop/docs/dev-database-reference-sql-vector-search).

## Trino and Spark

Trino and Spark insert with their own `INSERT` statements, and also copy query results with `CREATE TABLE ... AS SELECT`; Spark writes DataFrames too. See [ingest](/develop/docs/dev-database-guides-ingest#load-with-trino-spark-or-the-vast-query-engine).

See also: [import Parquet files](/develop/docs/dev-database-reference-sql-import-parquet), [ingest](/develop/docs/dev-database-guides-ingest), [UPDATE](/develop/docs/dev-database-reference-sql-update).