A Short Story About SQL’s Biggest Rival

Post by scott_meyer » Sat, Oct 10, 2020, 7:06 PM UTC

dreamcompiler wrote: post #93

tannhaeuser wrote: post #79

Had expected to read sth about Datalog, rivaling SQL at least in academic DB literature.

Same here. Datalog is superior to both, but alas.

Post by alquemist » Sun, Oct 11, 2020, 12:49 AM UTC

sergeykish wrote: post #117

alquemist wrote: post #116

Earlier quoted context omitted.

Assuming we key a and b by y.d and generalizing b to sets, we are reaching the limits of the relational model: there is no good way to represent multiple sets associated with a given key, we need independent tables to do so. Neither cross product nor null padding is a good way to represent the schema {y, a[], b[]}. y.d | a | b ------------------- y0 | a0 | b0 y0 | a0 | b1 y0 | a1 | b0 y0 | a1 | b1 y0 | a2 | b0 y0 | a…

# create table y (i int, d int); # insert into y values (1, 1), (1, 2), (2, 1); # select *, (select count(*) from (select count(y.i), y.d from y group by y.d) _) as foo from (select count(y.i), y.d from y group by y.d) _ ; count | d | foo -------+---+----- 1 | 2 | 2 2 | 1 | 2 (2 rows) By the way great example how unwieldy SQL is. A bit better with CTE: # with bar as (select count(y.i), y.d from y group by y.d) select…

Post by sergeykish » Sun, Oct 11, 2020, 8:29 AM UTC

alquemist wrote: post #123

sergeykish wrote: post #117

Earlier quoted context omitted.

# create table y (i int, d int); # insert into y values (1, 1), (1, 2), (2, 1); # select *, (select count(*) from (select count(y.i), y.d from y group by y.d) _) as foo from (select count(y.i), y.d from y group by y.d) _ ; count | d | foo -------+---+----- 1 | 2 | 2 2 | 1 | 2 (2 rows) By the way great example how unwieldy SQL is. A bit better with CTE: # with bar as (select count(y.i), y.d from y group by y.d) select…

Nice! Should be easy to do max instead of count. Then top instead of max. Then top3 breaks the model.

    # select * from
      (select count from (select count(y.i), y.d from y group by y.d) _ limit 2) b,
      (select count(y.i), y.d from y group by y.d) a;
     count | count | d 
    -------+-------+---
         1 |     1 | 2
         2 |     1 | 2
         1 |     2 | 1
         2 |     2 | 1
    (4 rows)

Post by scott_meyer » Sun, Oct 11, 2020, 4:58 PM UTC

refset wrote: post #122

scott_meyer wrote: post #121

Earlier quoted context omitted.

Don't despair. Datalog is alive and well. Yes, it does compose beautifully. However, with composition solved, you'll discover that a naive implementation of relational algebra will suffer from spurious cross products. https://engineering.linkedin.com/blog/2020/liquid-the-soul-o... https://engineering.linkedin.com/blog/2020/liquid--the-soul-...

This is pretty interesting stuff, thanks for sharing! Does temporal information get any special treatment in the economic graph at the moment? Are temporal queries of any interest? I ask because I work on Crux [0] which at its core is a point-in-time bitemporal Datalog engine, and I see a lot of similarities (schemaless core, Worst-Case Optimal Join etc.) [0] https://opencrux.com