Troubleshooting

The following are common issues that you may encounter while setting up the Knowledge Graph:

Many-to-Many Join Warning

If you are seeing this warning, it means the join between the first dataset can match many items in the second dataset, and vice-versa.

Why this is risky:

• In a many-to-many join, the result includes every combination of matches from both tables (A matches × B matches).

• That means the same record is repeated once for each match on the other table, which can overcount totals and distort aggregations e.g. averages, sums, etc.

• Unlike a one-to-many join (where only the “one” side repeats), both sides repeat here, so errors grow much faster and queries can get large and slow.

Example: AdClicks joined to Orders on user_id. A user can have many clicks and many orders, so the join returns every click×order pair. If a user has 5 clicks and 2 orders, the join yields 10 rows—so counting orders or summing order_amount is inflated by 5×.

What you can do:

• If you expected a single match, check your join fields are unique.

• If you only need totals, make a dataset that summarizes or deduplicates before joining.

• Add filters to limit matches.

If you want to move forward with the join anyway, we recommend adding a context node to the tables explaining the purpose of the join and the relationship between the two tables.