Comparing Trade Feeds to Detect Missing or Misordered Records

# Stock trading data across multiple vendors interview question


# Stock trading data across multiple vendors interview question












I had the following coding question in a quant shop interview recently. I have no experience with quant finance, so I was hoping to get some insight on if this problem actually represents some real world trading problem.

Essentially, the question was something like this:

> Suppose you have 2 vendors (can't remember if they used the term "vendor" or "exchange" but I don't think it matters for the problem) with different stocking trading data. Timestamps are given in the data. If the price and quantity are the same, then they're considered equal. This comparison can initially be done using a static comparison, and then using dynamic comparison. One followup was what if we allowed editing of the quantity and price? And another followup was adding additional vendors on top of the existing 2.

There was a huge language barrier issue, and I spent a lot of time trying to figure out what the interviewer was asking me to code, and I still don't quite understand it. What I remember is mentioned above, and I was hoping someone can tell me if is modeled after some real world trading problem?

## Answer by Bob Jansen (score 1)

https://quant.stackexchange.com/a/78205

I wasn't there so I can only guess but I think the below is reasonable:

- Exchanges is where trades trade and in this context data vendors are often third parties that deliver historical data on all kind of stuff used by the quants so this might be interview question material;

- It does happen that the data feeds don't exactly agree. Data can be missing or there can be other errors but you can certainly expect timestamps to differ if they record when the data is received by the vendor. The recorded order of events may change between vendors;

- Before using the data you want to know it's accurate;

- Therefore you could simply compare two trade feeds `trade1.price1 == trade2.price && trade1.quantity == trade2.quantity`. Maybe this was asked as a warm-up question;

- You could further improve the matching by checking whether different orders were recorded by looking at the previous or next trades in the two feeds. Maybe this was meant by dynamic.