Segmentation fault while using ThreadPoolExecutor about pypyodbc

[rolweber]

Could you please provide some example code that reproduces this problem?

[Ahmed1Arshed]

import pypyodbc as pypyod

conn = pyodbc.connect(
'DRIVER={ODBC Driver 17 for SQL Server};SERVER=******;DATABASE=;UID=;PWD=********'
)
cursor = conn.cursor()
update_data(row , config);
cursor.execute(
'UPDATE prices SET value = ?,freeze_expiry = ?,is_freeze = ?,pricing_method = ?,updated_by = ?,freeze_reason = ?,updated_at =CURRENT_TIMESTAMP WHERE id = ?',
(price, freeze_expiry, is_freez, pricing_method, updated_by,
freeze_reason, product_update_id))

executor = concurrent.futures.ThreadPoolExecutor(10)
for row in csv_file.itertuples(index=True):
executor.submit(update_data, row, config)

[rolweber]

Thank you. I'll leave it to @braian87b to decide whether that is a valid use case.

[Ahmed1Arshed]

what i am doing is reading a csv row which i need to update to my db. When i am using a single thread it's working smoothly but as i use Thread-pool it giving Segmentation fault (core dumped)

[rolweber]

What I see is a single DB connection, and a single cursor. But you're calling update_data from multiple threads, possibly simultaneously. I don't know whether that is supposed to work or not. If I had to guess, I'd say no.

[Ahmed1Arshed]

Why should i need to have multiple connection on every update.Connection is a simple thing that we are using that particular db for the operation we need to perform

[rolweber]

Because database connections are stateful?

And for the records: One thing that database connections are certainly not is simple.

[Ahmed1Arshed]

In that case it must be giving issue in single thread too isn't that so

[rolweber]

Database connections are certainly meant to execute multiple transactions in succession over the same connection. Creating a database connection is expensive, therefore they should be re-used in this fashion. Running multiple transactions over the same connection simultaneously is an entirely different matter.