---
metadata:
  - name: generator
    content: Diplodoc Platform v5.52.0
alternate:
  - https://ydb.tech/docs/en/yql/reference/syntax/upsert_into.md?version=v25.3
  - https://ydb.tech/docs/ru/yql/reference/syntax/upsert_into.md?version=v25.3
  - href: en/yql/reference/syntax/upsert_into.md
    type: text/markdown
    title: Markdown version
  - href: ../../../llms.txt
    type: text/markdown
    title: llms.txt
sourcePath: en/core/yql/reference/syntax/upsert_into.md
---
> **Documentation Index:** Fetch the complete configuration index at https://ydb.tech/docs/en/llms.txt

# UPSERT INTO

<!-- source: en/_includes/limitation-column-row-in-read-only-tx-warn.md -->
{% note warning %}

<!-- source: en/_includes/limitation-column-row-in-read-only-tx.md -->
Currently, mixing [column-oriented tables](https://ydb.tech/docs/en/concepts/glossary.md?version=v25.3#column-oriented-table) and [row-oriented tables](https://ydb.tech/docs/en/concepts/glossary.md?version=v25.3#row-oriented-table) in a single transaction is supported only if the transaction performs read operations; no writes are allowed. Support for read-write transactions involving both table types is under development.

If a write transaction includes both types of tables, it fails with the following error: `Write transactions that use both row-oriented and column-oriented tables are disabled at current time`.
<!-- endsource: en/_includes/limitation-column-row-in-read-only-tx.md -->

{% endnote %}
<!-- endsource: en/_includes/limitation-column-row-in-read-only-tx-warn.md -->

UPSERT (which stands for UPDATE or INSERT) updates or inserts multiple rows to a table based on a comparison by the primary key. Missing rows are added. For the existing rows, the values of the specified columns are updated, but the values of the other columns are preserved.



`UPSERT` and [`REPLACE`](https://ydb.tech/docs/en/yql/reference/syntax/replace_into.md?version=v25.3) are data modification operations that don't require a prefetch and run faster and cheaper than other operations because of that.

Column mapping when using `UPSERT INTO ... SELECT` is done by names. Use `AS` to fetch a column with the desired name in `SELECT`.

## Examples

```yql
UPSERT INTO my_table
SELECT pk_column, data_column1, col24 as data_column3 FROM other_table
```

```yql
UPSERT INTO my_table ( pk_column1, pk_column2, data_column2, data_column5 )
VALUES ( 1, 10, 'Some text', Date('2021-10-07')),
       ( 2, 10, 'Some text', Date('2021-10-08'))
```