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

# insert_or_update_records

Updates data in tables on databases connected with built-in PostgreSQL and SQL Connection. If the Primary Key does not exist, it will be registered as a new record.

```python morph_lib.database theme={"dark"}
# 📦 Package: morph_lib.database
# 🛠️ Function: insert_or_update_records

insert_or_update_records(
	data: pd.DataFrame,
	primary_keys: List[str]
	table_name: str,
	connection: Optional[str]
) -> None
```

対応データベース

* PostgreSQL
* MySQL
* Snowflake

### Parameters

<ParamField body="data" type="pd.DataFrame" required>
  The actual data to be passed to the function. The columns specified in primary\_keys must always be present.
</ParamField>

<ParamField body="primary_keys" type="List[str]" required>
  Specify the column names that will be the Primary Key of the table.
</ParamField>

<ParamField body="table_name" type="str" required>
  Specify the table name.
</ParamField>

<ParamField body="connection" type="str">
  The destination of the connection other than the builtin registered on the screen (optional)

  * If not specified, it will be set to built-in postgresql.
</ParamField>

### Example

```python theme={"dark"}
from morph_lib.database import insert_or_update_records

@morph.func
@morph.load_data("user_data")
def func_name(context):
	user_data: pd.DataFrame = context.data["user_data"]
	insert_or_update_records(
		user_data,
		["user_id"]
		"user",
	)
```