synapsesql

[woldea]

I have a python variable created in notebook cell and want to access it from SQL cell in the same notebook to write into a warehouse. Can anyone share me method to do this. Is there any method to write a python prepared sql code like upsert into a warehouse table using synapsesql? What I see now is we can write a dataframe int warehouse and read from warehouse into dataframe.

[ssrithar]

 Hi @woldea ,

This is a very common question in Microsoft Fabric notebooks, and the short answer is:

You cannot directly access a Python variable from a SQL cell.Python and SQL cells run in different execution engines.

However, there are supported patterns to achieve what you want, including UPSERT/MERGE into a Warehouse.

The most recommended one is  Write DataFrame → Warehouse (UPSERT supported)

If your Python variable can be represented as a DataFrame, this is the best practice.

Step 1: Python – prepare data

from pyspark.sql import Row

data = [
Row(id=1, name="Alice", amount=100),
Row(id=2, name="Bob", amount=200)
]

df = spark.createDataFrame(data)

Step 2: Write to Warehouse staging table

df.write \
.format("synapsesql") \
.mode("overwrite") \
.save("WarehouseName.dbo.stg_mydata")

Step 3: SQL cell – MERGE (UPSERT)

MERGE dbo.target_table AS tgt
USING dbo.stg_mydata AS src
ON tgt.id = src.id
WHEN MATCHED THEN
UPDATE SET
tgt.name = src.name,
tgt.amount = src.amount
WHEN NOT MATCHED THEN
INSERT (id, name, amount)
VALUES (src.id, src.name, src.amount);


This is the Fabric-supported UPSERT pattern.

If this post helps, then please appreciate giving a Kudos or accepting as a Solution to help the other members find it more quickly.
If I misunderstand your needs or you still have problems on it, please feel free to let us know. Thanks a lot!