test_diff.py

import pandas as pd

selected_columns = ["Amount", "Date", "Where"]


def find_diff_only(origin_df, target_df, identity="DBS"):
    shared_with_origin = pd.merge_asof(
        origin_df,
        target_df,
        on="Date",
        suffixes=("_origin", "_target"),
        direction="nearest",
        tolerance=pd.Timedelta(minutes=2),
    )
    origin_only = shared_with_origin[
        shared_with_origin["Amount_origin"] != shared_with_origin["Amount_target"]
    ]
    origin_only["Amount"] = origin_only["Amount_origin"]
    origin_only["Date"] = origin_only["Date"]
    origin_only["Where"] = identity
    origin_only = origin_only[selected_columns]
    for i in origin_only.index:
        origin_df.loc[i, "is_diff"] = 1
    return origin_only


def find_diff(dash_df, pos_df):
    dbs_only = find_diff_only(dash_df, pos_df, identity="DBS")
    pos_only = find_diff_only(pos_df, dash_df, identity="POS")
    final_diff = dbs_only.append(pos_only)
    return final_diff, len(final_diff), dbs_only.index, pos_only.index


dash_df = pd.DataFrame(
    {
        "Date": [
            pd.Timestamp("2020-03-25 13:30:00.023"),  # match
            pd.Timestamp("2020-03-26 13:30:00.023"),  # dash_only
            pd.Timestamp("2020-03-27 13:31:00.023"),  # match
            pd.Timestamp("2020-03-27 13:35:00.023"),  # dash_only
            pd.Timestamp("2020-03-27 13:35:00.023"),  # dash_only
        ],
        "Amount": [100, 200, 300, 400, 400],
    }
)

pos_df = pd.DataFrame(
    {
        "Date": [
            pd.Timestamp("2020-03-25 13:30:00.023"),  # match
            pd.Timestamp("2020-03-26 13:31:00.023"),  # pos_only
            pd.Timestamp("2020-03-27 13:32:00.023"),  # match
            pd.Timestamp("2020-03-27 13:38:00.023"),  # pos_only
        ],
        "Amount": [100, 210, 300, 400],
    }
)
find_diff(dash_df, pos_df)