diff_check.py

import pyodbc
import snowflake.connector

src = pyodbc.connect('DRIVER={ODBC Driver 17 for SQL Server};SERVER=<sql-server-host>;DATABASE=<source-database>;Trusted_Connection=yes')
tgt = snowflake.connector.connect(
    account='<snowflake-account>', user='<user@domain>',
    authenticator='externalbrowser', warehouse='<snowflake-warehouse>',
    database='<snowflake-database>', schema='<target-schema>', role='<snowflake-role>'
)

tables = [
    ('SOURCE_TABLE_A', 'TARGET_TABLE_A', 'id'),
    ('SOURCE_TABLE_B', 'TARGET_TABLE_B', 'name'),
    ('SOURCE_TABLE_C', 'TARGET_TABLE_C', 'category'),
]

for src_tbl, tgt_tbl, order_col in tables:
    print(f"\n{'='*60}")
    print(f"TABLE: {src_tbl}")

    sc = src.cursor()
    sc.execute(f"SELECT TOP 2 * FROM {src_tbl} ORDER BY [{order_col}]")
    src_cols = [d[0] for d in sc.description]
    src_rows = sc.fetchall()
    print(f"SOURCE cols: {src_cols}")
    for r in src_rows:
        print(f"  SRC: {list(r)}")

    tc = tgt.cursor()
    tc.execute(f'SELECT * FROM <snowflake-database>.<target-schema>.{tgt_tbl} ORDER BY {order_col} LIMIT 2')
    tgt_cols = [d[0] for d in tc.description]
    tgt_rows = tc.fetchall()
    print(f"TARGET cols: {tgt_cols}")
    for r in tgt_rows:
        print(f"  TGT: {list(r)}")

    src_set = set(c.upper() for c in src_cols)
    tgt_set = set(c.upper() for c in tgt_cols)
    if src_set != tgt_set:
        print(f"  EXTRA in target:   {tgt_set - src_set}")
        print(f"  MISSING in target: {src_set - tgt_set}")

src.close()
tgt.close()
print("\nDONE")