CQL - Csv Query Language

CQL is a a query language for performing operations on csv files. The language is verry simmular to ordanary SQL. The language have two query types Select and Create. I choose to not implement Update and Delete since their is no way to ensure the data integrety in a csv file, which you can in a SQL table. The parser is written in F# with Fparsec and supports recursive grammer.

Exampel - select query

The following query performes an inner join on the files person.txt and adress.txt and then filter out the person above 18 which live in Sweden or in Germany. The result is orderd by age and the top 3 rows are selected.

select '*' from 'person.txt' as 'p' 
inner join 'adress.txt' as 'a' 'a.ownerid' = 'p.userid' 
where 2021 - 'p.birthyear' > 18 
where 'a.country' = "Sweden" || 'a.country' = "Germany" 
order by 'p.birthyear' asc 
top 3

This is a representation of the files person.txt and adress.txt

person.txt                                        adress.txt
.........................................         ...................................................
userid; | name;     |  age;   | gender  |         | ownerid;  | country;  | city;       | zip       |
.........................................         ..................................................|
928;    | Lars;     | 1967;   | male    |         | 1394;     | Germany;  | Berlin;     | 43-242    |
1394;   | Klaus;    | 1989;   | male    |         | 928;      | Sweden;   | Stockholm;  | 97-1235   |
571;    | Juliette; | 1982;   | female  |         | 928;      | Norway;   | Oslo;       | 324-2425  |
352;    | Moa;      | 1993;   | female  |         | 571;      | France;   | Paris;      | 921-135   |
25;     | Teo;      | 2008;   | male    |         | 352;      | Sweden;   | Karlstad;   | 456-123   |
82;     | Olavi;    | 2006;   | male    |         | 25;       | Sweden;   | Gothenburg; | 10-412    |
122;    | Lena;     | 1996;   | female  |         | 82;       | Finland;  | Helsinki;   | 789-1     |
.........................................         | 122;      | Germany;  | Hamburg;    | 23-1346   |
                                                  ...................................................

The exampel query will get parsed into a domain model which looks like this

Select
  ([All], TableName ("person.txt", Some "p"),
   [Inner
      (TableName ("adress.txt", Some "a"),
       EqualityExpression
         (Equals, ColumnIdentifier (Some "a", "ownerid"),
          ColumnIdentifier (Some "p", "userid")))],
   [Where
      (EqualityExpression
         (GreaterThan,
          ArithmeticExpression
            (Subtract, Litteral (NumericLitteral 2021.0),
             ColumnIdentifier (Some "p", "birthyear")),
          Litteral (NumericLitteral 18.0)));
    Where
      (LogicalExpression
         (Or,
          EqualityExpression
            (Equals, ColumnIdentifier (Some "a", "country"),
             Litteral (StringLitteral "Sweden")),
          EqualityExpression
            (Equals, ColumnIdentifier (Some "a", "country"),
             Litteral (StringLitteral "Germany"))))],
   Some ((Some "p", "birthyear"), Ascending), 
   Some (Top 3))

The domain model will then get evaluated and result in the following table.

userid   name   birthyear   gender   ownerid  country   city        zip
   928   Lars   1967        male     928      Sweden    Stockholm   97-1235
   1394  Klaus  1989        male     1394     Germany   Berlin      43-242
   352   Moa    1993        female   352      Sweden    Karlstad    456-123