PigletQL - a tiny SQL-like interpreter

PigletQL is a tiny SQL-like language featuring a single value type, three kinds of queries (CREATE TABLE, INSERT, SELECT) and a Volcano-style (see the “Volcano - an extensible and parallel query evaluation system” paper) interpreter. It does not do any optimization and keeps all data in memory.

It was created as an example language for an article (Russian, English) on SQL interpreter architecture.

Building

It should be trivial to compile the interpreter on Linux using a recent GCC. Just clone the repo and do:

> make
> make test
> ./pigletql
> # your query here

Example queries

> ./pigletql
> create table rel1 (a1,a2,a3);
> insert into rel1 values (1,2,3);
> insert into rel1 values (4,5,6);
> select a1 from rel1;
a1
1
4
rows: 2
>

> ./pigletql
> create table rel1 (a1,a2,a3);
> insert into rel1 values (1,2,3);
> insert into rel1 values (4,5,6);
> select a1 from rel1 where a1 > 3;
a1
4
rows: 1
>

> ./pigletql
> create table rel1 (a1,a2,a3);
> insert into rel1 values (1,2,3);
> insert into rel1 values (4,5,6);
> select a1 from rel1 order by a1 desc;
a1
4
1
rows: 2

> ./pigletql
> create table rel1 (a1,a2,a3);
> insert into rel1 values (1,2,3);
> insert into rel1 values (4,5,6);
> create table rel2 (a4,a5,a6);
> insert into rel2 values (7,8,6);
> insert into rel2 values (9,10,6);
> select a1,a2,a3,a4,a5,a6 from rel1, rel2 where a3=a6;
a1 a2 a3 a4 a5 a6
4 5 6 7 8 6
4 5 6 9 10 6
rows: 2

Code structure

• pigletql-eval.h - evaluation engine, i.e. Volcano-style operators, relations, tuples, etc
• pigletql-parser.h - lexer/parser
• pigletql-validate.h - query validation logic
• pigletql-catalogue.h - a catalogue of relations available
• pigletql-def.h - constants and helpers
• pigletql.c - putting everything together