> For the complete documentation index, see [llms.txt](https://docs.bdb.ai/data-pipeline/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.bdb.ai/data-pipeline/~/changes/BozVtU3tL5BwgOHeoMZt/components/writers/db-writer.md).

# DB Writer

All component configurations are classified broadly into the following sections:

* ​[Basic ](https://docs.bdb.ai/7.6/data-pipeline/components/component-base-configuration)​
* Metadata
* ​[Resource Configuration](https://docs.bdb.ai/7.6/data-pipeline/components/resource-configuration)​
* ​[Connection Validation](https://docs.bdb.ai/7.6/data-pipeline/components/connection-validation)​

{% hint style="success" %}
*Please check out the given demonstration to configure the component.*
{% endhint %}

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fuq3RSHHup7fjHYaspk7y%2Fuploads%2F3fKUwcnQmIJIVu7uA5zY%2Fdb%20writer%20(online-video-cutter.com).mp4?alt=media&token=8817a387-b4b5-4a13-9e52-db39aa0a3a5c>" %}
Configuring the DB Writer Component as a part of Pipeline Workflow
{% endembed %}

### **Drivers Available**

* MySQL
* Oracle
* PostgreSQL
* MS-SQL
* ClickHouse
* Snowflake

{% hint style="info" %}
*<mark style="color:green;">Please Note:</mark>*&#x20;

* *The ClickHouse driver in the Spark components will use HTTP Port and not the TCP port.*
* It is always recommended to create the table before activating the pipeline to avoid errors as RDBMS has a strict schema and can result in errors.
  {% endhint %}

### Save Modes <a href="#save-modes" id="save-modes"></a>

The RDBMS writer supports 3 save modes:

#### **Append**

As the name suggests it adds all the records without any validations.

#### **Overwrite**

This mode truncates the table and adds fresh records. after every run you will get records that are part of the batch process.

#### **Upsert**

This operation allows the users to insert a new record or update existing data into a table. For configuring this we need to provide the Composite Key.

The BDB Data Pipeline supports composite key based upsert, in case of composite key, we can specify the second key by using comma separator e.g., *key1, key2​.* It has now an option to **upload the spark schema.** This can greatly improve the speed of the write operation as the component will ignore inferring schema and go with the provided schema.

<figure><img src="https://363587200-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fuq3RSHHup7fjHYaspk7y%2Fuploads%2FMYhU9zeZuo2UmVF4tnSa%2FMicrosoftTeams-image%20(67).png?alt=media&amp;token=1008a27a-1947-4763-90f3-6879d9805240" alt=""><figcaption><p>Spark Schema upload</p></figcaption></figure>

{% hint style="info" %}
*<mark style="color:green;">Please Note</mark>: For ClickHouse Component Upsert is comparatively slow. It is preferable to create a table where the engine is **ReplacingMergeTree** and a view where we load the view with the Final clause. **In the component keep the write mode to Append**.*
{% endhint %}

* ***Query:*** In this field, we can write a DDL for creating the table in database where the in-event data has to be written. For example, please refer the below image:

<figure><img src="https://363587200-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fuq3RSHHup7fjHYaspk7y%2Fuploads%2FPvQvOQwIZG5eKJ5E0gS4%2FMicrosoftTeams-image%20(68).png?alt=media&amp;token=8dafef5d-a518-402b-84d3-bcac3c35aa69" alt=""><figcaption></figcaption></figure>