Python Forum

I'm trying to load a PostgreSQL table with information from MSsql. I have 2 tables that are identical on both db's. On the Mssql side there is a column called ImageSource which is defined as Image, and on the PG side I have it as Bytea. The script posted is just trying to pull 1 record to test with, but I'm running into an issue in the DF portion. The 2 tables have 21 fields defined.

This is the error message I'm receiving. I can see it connect, and my column list shows 21 fields. I blanked out all db info stuff but that connection stuff is working as I can see activity(connections) on both DB's

When I get this working, I'll have 50,000 to 100K records to export daily so if you see any speed improvements I can make please denote as I'm very new to Python. hence the BatchSize

Error Message from --> df = pd.DataFrame(batch, columns=columns)

raise ValueError(f"Shape of passed values is {passed}, indices imply {implied}")
ValueError: Shape of passed values is (1, 1), indices imply (1, 21)

How do I fix this error?

Many thanks.

import pyodbc
import psycopg2
import io
from sqlalchemy import create_engine
import pandas as pd

import time
from tqdm import tqdm

import numpy as np

# MSSQL connection
mssql_conn = pyodbc.connect('DRIVER={SQL Server};SERVER=xxxx;DATABASE=yyyy;UID=xxxx;PWD=xxxx;Trusted_Connection=yes')
mssql_cursor = mssql_conn.cursor()

# PostgreSQL connection
pg_conn = psycopg2.connect(database="hhhh", user="xxxxx", password="xxxxx", host="xxxxx", port="xxxx")
pg_cursor = pg_conn.cursor()

# Create SQLAlchemy engine for PostgreSQL
engine = create_engine('postgresql://postgres:xxxxx@xxxxx/xxxxx')

# Query to fetch data from MSSQL
mssql_query = "select id,imagename,imagetype,imagesource,receiveddatetime,imagepath,site,machinenbr,linenbr,takeupnbr,spoolnbr,imageindex,cameranbr,spoolstartdt,left((convert(time(0),SpoolStartTime)),8) as SpoolStartTime,DefectDate,left((convert(time(0),DefectTime)),8) as DefectTime,defectnbr,defectclass,reviewer,UserDefectInput FROM Image_Classification_Master(nolock) WHERE id = 13604630"
mssql_cursor.execute(mssql_query)

# Fetch column names
columns = [column[0] for column in mssql_cursor.description]

# Process and insert data in batches
batch_size = 5000
while True:
    batch = mssql_cursor.fetchmany(batch_size)
    if not batch:
        break

    # Convert batch to DataFrame
    df = pd.DataFrame(batch, columns=columns)

    

    # Process Image field
    if 'ImageSource' in df.columns:
        df['ImageSource'] = df['ImageSource'].apply(lambda x: psycopg2.Binary(x) if x else None)

    # Insert batch into PostgreSQL
    df.to_sql('Image_Classification_Master', engine, if_exists='append', index=False, method='multi', chunksize=batch_size)

    total_rows = len(df)
    rows_inserted = 0

    start_time = time.time()

with tqdm(total=total_rows, desc="Inserting data") as pbar:
    for i in range(0, total_rows, batch_size):
        batch = df.iloc[i:i+batch_size]
        batch.to_sql('Image_Classification_Master', engine, if_exists='append', index=False, method='multi')

        rows_inserted += len(batch)
        pbar.update(len(batch))

        # Calculate and display progress
        progress = (rows_inserted / total_rows) * 100
        elapsed_time = time.time() - start_time
        estimated_total_time = (elapsed_time / rows_inserted) * total_rows
        remaining_time = estimated_total_time - elapsed_time

        print(f"Progress: {progress:.2f}% | Rows inserted: {rows_inserted}/{total_rows}")
        print(f"Elapsed time: {elapsed_time:.2f}s | Estimated time remaining: {remaining_time:.2f}s")

print("Insertion complete!")

# Close connections
mssql_cursor.close()
mssql_conn.close()
pg_cursor.close()
pg_conn.close()

I changed the df and reshape.

reshaped_data = np.array(batch).reshape(1, 21)
    df = pd.DataFrame(reshaped_data, columns=['id', 'imageName', 'imageType', 'imageSource', 'receiveddatetime',
       'imagepath', 'site', 'machinenbr', 'linenbr', 'takeupnbr', 'spoolnbr',
       'imageindex', 'cameranbr', 'spoolstartdt', 'spoolstarttime',
       'defectdate', 'defecttime', 'defectnbr', 'defectclass', 'reviewer',
       'userdefectinput'])

I also modified query to get a single record reshape(1, 21) for testing .. now I get this error:

Error:psycopg2.errors.UndefinedColumn: column "id" of relation "Image_Classification_Master" does not exist
LINE 1: INSERT INTO "Image_Classification_Master" (id, "imageName", .

The id column does exists in both mssql and PG db's

hartman60

hartman60