3 ms·
I did something similar for a NoSQL database [1]. The biggest surprise was how much the query performance can change for an index when the data distribution cha
by realPubkey 4y ago
I did something similar for a NoSQL database [1].
The biggest surprise was how much the query performance can change for an index when the data distribution changes slightly. For example using a real distribution for an 'age' field instead of just using a random number like in the test data.
[1] https://rxdb.info/query-optimizer.html https://rxdb.info/query-optimizer.html
- HPsquared 4y agoThe 'age' thing might be due to the difference between floats and ints? Integer "age" has many repeats, but random floats are unique. That, or random ints might be from a large pool, again not many repeats.
- contingencies 4y agoSin #5 - Unrepresentative workloads: The common assumption in academic research systems is that the cluster workload is relatively homogenous. Most research evaluations measure performance by running a single job on an otherwise idle cluster. - Schwarzkopf et al, 'The seven deadly sins of cloud computing research' (2014) .. via https://github.com/globalcitizen/taoup https://github.com/globalcitizen/taoup
- jessaustin 4y agoIt seems better for birthdate to be stored in the database and age just to be calculated when needed?
- realPubkey 4y agoYes of course you better store the birthday normally. This was just a projection of the real data set, used in my testings.