Prakhar Shrivastava

Join regional orders using the complete source key. Test how reused order IDs mix customers and inflate a $240 total to $420 in a runnable SQL example.

Test how pandas ffill can copy values between accounts. Preserve entity boundaries, sort by time, and distinguish missing updates from unknown values.

Test why EXISTS with COUNT(*) can include customers with no orders, then use matching rows or an explicit count comparison.

Test pandas merge_asof with future prices, exact timestamps, and stale records. Use explicit matching rules and preserve the reference timestamp.

Count distinct account–campaign pairs with tested SQL, avoid concatenation collisions, and report unresolved identifiers separately.

Distinguish missing revenue, known subtotals, and genuine zeros in pandas. Test min_count=1 alongside row counts and explicit completeness checks.

Fix a revenue-share denominator changed by WHERE. Run a tested SQL example separating population filters from display filters, including zero totals.

Compare PERCENT_RANK and CUME_DIST with a tested SQL spend example covering ties, all-equal groups, single customers, and dashboard thresholds.

Audit duplicate order keys in pandas, separate identical replays from conflicting amounts, and understand keep=False with a tested USD order example.

Debug NaN values after assigning a pandas Series. Test index alignment, keyed lookup, and a positional assignment that swaps amounts while preserving totals.