How to handle INT64 values from Parquet when copying to a DateTime column in Azure Synapse using Copy Activity?

Question

How to handle INT64 values from Parquet when copying to a DateTime column in Azure Synapse using Copy Activity?

HARISH LOGANATHAN 20

I have a Parquet file stored in ADLS. One of the columns in the Parquet file is stored as INT64.

When I try to copy this data from ADLS into an Azure Synapse table using Azure Synapse Analytics Copy Activity, I get the following error:

"errorCode": "2200",
"message": "ErrorCode=TypeConversionNotSupported Data Types,
Exception occurred when converting value '1768842423000000'
for column name 'UP_DATE' from type 'Int64' (precision:-1, scale:-1)
to type 'DateTime' (precision:255, scale:7). Additional info:",
"failureType": "UserError",
"target": "COPY CDZ to DDZ",
"details": []

The target Synapse table column is of type DateTime, but the Parquet file contains the value as Int64.

Is there any configuration in Copy Activity that can automatically convert this value into a proper DateTime format, or any recommended workaround?

Answer accepted by question author

0 additional answers

Your answer

Answer 1

Smaran Thoomu 33,920 Microsoft External Staff Moderator

Hi @HARISH LOGANATHAN
The error happens because Synapse Copy Activity cannot automatically convert an INT64 value into a DateTime column.

The value 1768842423000000 looks like a Unix timestamp stored in microseconds. Copy Activity does not automatically interpret that as a date.

There is no direct setting in Copy Activity to convert INT64 → DateTime automatically.

Here are the recommended options:

Option 1 – Use Mapping Data Flow

Instead of Copy Activity, use a Mapping Data Flow and add a derived column with conversion logic like:

If value is in microseconds → divide by 1,000,000
Convert it using toTimestamp()

Example logic:

toTimestamp(UP_DATE / 1000000)

This allows proper conversion before writing to Synapse.

Option 2 – Stage as BIGINT first

Create the target column in Synapse as BIGINT.
Load the data as-is.
Use a SQL update statement to convert it:

Example:

UPDATE table

Then change the column type if needed.

Option 3 – Convert before landing

If possible, convert the column to proper timestamp format before writing the Parquet file.

Hope this helps. Let me know if you have any questions.

HARISH LOGANATHAN 20 Reputation points

2026-03-02T12:35:52.2033333+00:00

For option 3, could you please suggest proper timestamp format.
Smaran Thoomu 33,920 Reputation points Microsoft External Staff Moderator

2026-03-03T05:48:55.57+00:00
HARISH LOGANATHAN For option 3, the proper timestamp format depends on how you are generating the Parquet file.

If possible, the best approach is to store the column as a true timestamp type instead of INT64.

For example:

In Spark / Databricks, cast the column to timestamp before writing:

col("UP_DATE").cast("timestamp")

Or convert your Unix microseconds value like this:

to_timestamp(UP_DATE / 1000000)

When written to Parquet as a proper timestamp type (TIMESTAMP_MILLIS or TIMESTAMP_MICROS), Synapse Copy Activity can correctly map it to a DateTime column.

If you are generating the file outside Spark, the value should be stored in standard ISO 8601 format, such as:

2026-02-28 14:35:12

or

2026-02-28T14:35:12Z

So the key point is: Instead of storing it as raw INT64, convert it to a proper timestamp data type before writing the Parquet file.

Let me know what tool you are using to generate the Parquet file, and I can give a more specific example.

Share via

How to handle INT64 values from Parquet when copying to a DateTime column in Azure Synapse using Copy Activity?

0 additional answers

Your answer