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

# FAQ

> Answers for the current PostgreSQL-to-PostgreSQL MantrixFlow workflow.

## Which connectors are live in the public docs?

PostgreSQL source and PostgreSQL destination.

Other connector pages may exist as direct-link notices, but they are not documented as live self-serve workflows.

## Does MantrixFlow create destination tables?

No. Create the destination table before running the pipeline. The runner validates the table and columns, then fails with a clear error if the destination is missing or mismatched.

## Where do I select source tables?

Open the builder and click the settings icon on the **Source** node. Use **Discover schema**, **Refresh tables**, **Include**, and **Preview**.

## Where do I write SQL?

Use the **SQL Layer** tab in the **Destination** panel.

Reference staged source data with the source helper:

```sql theme={"theme":{"light":"github-light","dark":"github-dark"}}
SELECT *
FROM {{ source('raw', 'public__orders') }}
```

## What Normalisation rules are available?

The live Normalisation rules are Rename and Cast. They run before the SQL Layer.

## Which write mode should I use?

Use **Upsert**. It is the live write mode in the current app.

If the destination table has a primary key, delivery uses merge behavior. If it has no primary key, delivery appends rows and records a no-primary-key warning.

## Can I schedule a Full Table sync?

No. Scheduling is enabled for Incremental sync. For Full Table sync, run manually.

## Why did my run fail before writing rows?

Common causes:

* source table does not exist or permission was removed
* destination table does not exist
* SQL model output contains a column missing from the destination table
* Incremental sync is missing a replication key
* credentials changed after the connection was saved

## Is staging data permanent?

No. Runs stage data in temporary storage. The runner extracts checkpoint state and cleans up the staging file during cleanup.
up.