How do nested MongoDB documents land in PostgreSQL?

In Weld's unpacked mode, each top-level field becomes a column and nested objects one level down are kept as JSON in that column. In packed mode, each row holds the id and the full document as JSON. You choose per connection and can change it, which triggers a full re-sync.

Is the MongoDB to PostgreSQL sync incremental?

Yes. Weld syncs MongoDB in incremental merge mode using _id as the cursor by default, which picks up new documents. To capture updates as well, set the cursor to a timestamp field that changes on modification.

Does Weld support MongoDB change streams or CDC?

Not yet. Weld's docs list MongoDB CDC as under development. Today, collections sync on a schedule using a cursor field.

Can I sync a MongoDB sharded cluster to PostgreSQL?

Yes. Weld has a separate MongoDB Sharded Cluster connector with the same packed and unpacked modes.

Can AWS DMS replicate MongoDB to PostgreSQL?

Yes. AWS DMS supports MongoDB as a source with document mode, which loads each document into a _doc column, or table mode, which flattens nested fields into dot-named columns. Ongoing replication reads the oplog, so the source must run as a replica set.

MongoDB to PostgreSQL

Turn MongoDB collections into PostgreSQL tables you can query with SQL. Weld lands each collection as a table, with top-level fields as columns or the whole document as JSON, and keeps it current with incremental syncs.

Start for free View setup guide

4.8 on G2

Source

Weld

Warehouse

Checked against Weld and vendor docs on 10 October 2026

Weld · New sync

SourceMongoDBConnected

Tables to sync

Your own tables, picked when you set up the sync.

Extract
Load
Keep in sync

DestinationPostgreSQLReady

mongodb schema

mongodb.Synced
mongodb.Syncing
mongodb.Synced
mongodb.Synced

Trusted by data and AI teams

Options

Three ways to get MongoDB data into PostgreSQL

Going from documents to rows means deciding what happens to nested fields. Each method below answers that differently, and that choice shapes every query you write afterwards.

mongoexport and a load script

Best for: A one-off export of a few collections
Effort: Medium: export, flatten, create tables, load
Freshness: Whenever you rerun it
Good to know: Exports Extended JSON or CSV; CSV needs an explicit field list. MongoDB says it is not a backup tool. Flattening nested documents and typing columns is up to your script.

AWS Database Migration Service

Best for: Teams on AWS who want full load plus CDC
Effort: Medium: replica set, endpoints, mode choice, tasks
Freshness: Full load, then ongoing replication
Good to know: Document mode puts each document in a _doc column; table mode flattens nested fields into dot-named columns from a scan of 1,000 documents by default. CDC needs a replica set for the oplog.

Weld

Best for: Querying MongoDB data next to other sources in PostgreSQL
Effort: Low: connection string, pick collections, choose pack mode
Freshness: Incremental syncs on your schedule
Good to know: Unpacked mode makes top-level fields columns and keeps nested objects as JSON; packed mode lands id plus the whole document. Merge syncs use _id by default.

How it works

How to sync MongoDB to PostgreSQL with Weld.
In the app, with the API or an agent.

Setup guide API guide MCP docs

Weld · Data warehouse

Data warehousePostgreSQLConnected

Data locationChosen when you connect

CredentialsService account or key, stored encrypted

Data

What lands in PostgreSQL
From MongoDB to PostgreSQL.

MongoDB has no fixed schema, so there is no list of tables to publish. Each collection you select becomes one PostgreSQL table, in the shape your pack mode decides.

Unpacked collection tables
One column per top-level field, including _id, with data types inferred. A nested object such as nested: { name: 3 } lands as JSON in a single nested column.
Packed collection tables
Two columns: id and json, the full document. Useful when documents in a collection vary a lot and you prefer to parse fields in SQL.
_weld_synced
The time Weld synced each row. Use it to check freshness and to build incremental models on top.

Example query

Does each collection have one row per document, and how fresh is it?
In PostgreSQL SQL.

Runs in PostgreSQL against a collection synced in unpacked mode, assuming you named the destination schema mongodb. Replace orders with one of your collections. _id comes from MongoDB and _weld_synced from Weld.

PostgreSQL · query

1select2  count(*) as row_count,3  count(distinct "_id") as documents,4  max(_weld_synced) as last_synced5from mongodb."orders"

Weld MCP server

Ask your agent.
It queries PostgreSQL for you.

Connect Claude, Cursor, ChatGPT or your own agent to Weld over MCP. It reads the MongoDB context, finds the tables it needs and runs the query in PostgreSQL.

Set up the MCP server

Claude CodeOpus 5

ConnectedWeld MCP

Which MongoDB tables do we have in PostgreSQL, and which is the largest?

Looking up what Weld has synced from MongoDB.

●weld·search_available_references

⎿raw.mongodb.*

tables, columns and freshness

●weld·get_reference_schema

⎿columns, types and nullability

●weld·run_query

⎿select count(*) from {{raw.mongodb.…}}

submitted

●weld·get_query_status

⎿DONE

●weld·get_query_result

⎿rows returned

✓Answered from MongoDB data in PostgreSQL

>Ask about your MongoDB dataMCP

Under the hood

Things to know.
Before you sync MongoDB.

Nested documents stay JSON below the first level: Unpacked mode unpacks one layer of nested fields. Anything deeper stays as JSON in its top-level column, so plan to extract nested values in SQL models rather than expecting one column per path.
The default _id cursor only sees new documents: Syncs run in incremental merge mode on _id by default, and _id reflects creation time, so updates to existing documents are missed. Set the cursor to a field that records modification time, and index it in MongoDB.
Switching pack mode re-syncs the table: Changing a collection between packed and unpacked mode triggers a full re-sync of that table, and its columns change shape. Pick the mode before you build models on it.
Deletes are not covered by scheduled syncs: Weld's MongoDB docs do not describe delete handling for incremental syncs. MongoDB CDC, which streams document deletes, is listed as coming soon, so check deleted documents yourself if they matter to your reports.

Sources: Weld MongoDB connector docs · Weld MongoDB CDC docs · Weld MongoDB Sharded Cluster docs · Weld PostgreSQL destination docs · Weld managed fields · MongoDB: mongoexport · AWS DMS: Using MongoDB as a source