Extract table name and columns from SQL schema - regex

Question

I need help to export table-name & columns from table schema (DDL) using regex.
CREATE TABLE todos (
id INTEGER NOT NULL,
user_id INTEGER NOT NULL,
team_id INTEGER NOT NULL,
title TEXT NOT NULL DEFAULT "Hello World!",
description TEXT NOT NULL UNIQUE,
UNIQUE (title),
PRIMARY KEY (id),
FOREIGN KEY (user_id) REFERENCES users (id),
FOREIGN KEY (team_id) REFERENCES teams (t_id)
ON UPDATE RESTRICT
ON DELETE RESTRICT
)
Table name
todos
2. Columns
id // as group 1 (column name)
INTEGER // as group 2 (column type)
NOT NULL // as group 3 (column nullable) empty if nothing
DEFAULT // as group 4 (default value for example "Hello World")
UNIQUE // as group 5 (column uniqueable) empty if nothing
Note: UNIQUE can be also on table level same as title column.
3. Primary key
id // as group 1 (primary key)
Table level: PRIMARY\sKEY\s+\(([^\)]+)\)
Column level: check below answer.
4. Foreign keys:
// first
user_id // as group 1 (foreign key)
users // as group 2 (reference table name)
id // as group 3 (reference primary)
// second
team_id // as group 1 (foreign key)
teams // as group 2 (reference table name)
t_id // as group 3 (reference primary)
ON UPDATE RESTRICT // as group 4
ON DELETE RESTRICT // as group 5
I've found a simple regex in [github] (https://github.com/yiisoft/yii2/issues/6351#issuecomment-91064631) but not support RESTRICT
/FOREIGN KEY\s+\(([^\)]+)\)\s+REFERENCES\s+([^\(^\s]+)\s*\(([^\)]+)\)/mi

Answer 1

Extract a table name:
CREATE\s+TABLE\s+([\w_]+)
Get column names:
\s+([\w_]+)[\s\w]+,
Get a primary key field:
\s*PRIMARY\s+KEY\s+\(([\w_]+)\)
Get foreign keys data:
\s*FOREIGN\s+KEY\s+\(([\w_]+)\)\s+REFERENCES\s+([\w_]+)\s+\(([\w_]+)\)
You can test it here (respectively):
https://regexr.com/59251
https://regexr.com/59254
https://regexr.com/5925a
https://regexr.com/594eb

Answer 2

The Regex are returning results into a named captured group, you can find the name if you look here (?'GREOUP-NAME'..myregex...). It makes it easier for you to reference them after a finished regex search, it will be easier to split them.
FULL SEARCH
((?'COLUMN_NAME'(?<=^\s\s)([[:lower:]]\w+))|(?'PRIMARY_KEY'(?<=PRIMARY\sKEY\s\()(\w+))|(?'TABLE_NAME'(?<=\bTABLE\s)(\w+)))
SPLIT SEARCH
Get table name:
(?'TABLE_NAME'(?<=\bTABLE\s)(\w+))
Get primary key:
(?'PRIMARY_KEY'(?<=PRIMARY\sKEY\s\()(\w+))
Get column name: This one is a little bit sloppy and will only capture columns that are lowercase. Since your text didn't have any tabs-characters. This was the best i could do but it's a bit risky.
(?'COLUMN_NAME'(?<=^\s\s)([[:lower:]]\w+))
You can run them here, regex101, and try it out.
Be aware that the regex is dependent on whatever regex-engine your are using. There are some shortcomings regarding standards, and some regex's might need to be translated to your engine. For ex. lookbehind is not supported on all engines.