28. How to Merge and Join Tables in pandas for Blinkit Data

28.6 Common Mistakes (Blinkit Flavour)

By Ravindra Bagale, trainer LinkedIn Last updated: 24 October 2026

Why. Same bugs as Chapter 23, but with store codes like PUN-01 vs PUN01 and rider IDs with spaces.

What.

Key typos / strip issues ("PUN-01 ").
Many-to-many duplicate keys → row explosion → double GMV.
Merging on city name only (two stores in Pune collide).
Forgetting how (default is inner!).
concat when you meant merge (stack vs join).

How.

Hardening

stores["store_id"] = stores["store_id"].astype(str).str.strip().
Prefer stable IDs over display names for keys.
Always pass how= explicitly.
assert stores["store_id"].is_unique before left-joining orders.
Use concat only when stacking the same grain (monthly files) — Chapter 8.1.

Worked example A — Default how trap.

import pandas as pd
a = pd.DataFrame({"store_id": ["PUN-01", "X"], "amount_inr": [1, 2]})
b = pd.DataFrame({"store_id": ["PUN-01"], "city": ["Pune"]})
print("default", len(a.merge(b, on="store_id")))  # inner -> drops X
print("left", len(a.merge(b, on="store_id", how="left")))

Worked example B — City-only key failure.

Two Pune stores merged on city duplicate every Pune order — use store_id.

Worked example C — concat reminder.

sep = pd.DataFrame({"order_id": ["A"], "city": ["Pune"]})
oct_ = pd.DataFrame({"order_id": ["B"], "city": ["Nashik"]})
print(pd.concat([sep, oct_], ignore_index=True))  # stack months — not a store join

Ravindra Bagale's Tip

how= explicit. strip keys. unique dimension. GMV double = explosion alarm.