LocustDB

LocustDB copied to clipboard

In this function you use the unsafe keyword for some safe expressions. We need to mark unsafe operations more precisely using unsafe keyword. Keeping unsafe blocks small can bring many...

cactter

cswinter

Why not support ansi SQL?

1

the speed is great, but have to learn special SQL grammar.

alitrack

LocustDB currently doesn't support explicit GROUP BY clauses, and instead implicitly performs GROUP BY of any non-aggregation expression if your query contains an aggregation expression. So e.g. `SELECT COUNT(1), page,...

cswinter

Default ordering of index columns and inserts to the already existing data

1

Do you have any thoughts on sparse indexes and column preordering (like it is done in the Clickhouse, for instance)? I haven't found anything related (maybe it's already there, sorry...

ravlio

Support window functions like row_number()

4

This would be nice to have.

jrmuizel

Queries that have type errors or reference missing columns should give helpful errors/warnings

7

STR: - load https://flatteningthecurve.herokuapp.com/data/canadatesting - execute `select cumulative_testing from default where province = "Ontario"` - you get: `Type error: Function Equals is not implemented for types Type { decoded: String,...

jrmuizel

Order by string column fails with `top_n_asc not supported for type ScalarStr`

4

With https://flatteningthecurve.herokuapp.com/data/canadatesting ``` $ ./repl --load /tmp/canada_testing.csv locustdb> select * from default where province = 'Ontario' order by 'date' Some assumption was violated. This is a bug: top_n_asc not supported...

jrmuizel

Revisit choice of hash functions during hash grouping

1

Currently everything uses FNV, but there may be better choices in particular for string data. Should set up some benchmarks and find out.

cswinter

`count_by_passenger_count_pickup_year_trip_distance` is 50% slower after query refactor.

cswinter