如何将sqlite3数据库中的所有表合并为一个，但仅包含唯一内容？ （不包括相同的行数据）

Question

FOr example, I have 4 tables, and in table 1 there is a row with the same data as a row in table 2. In table 3 there is a row that looks the same as a row in table 4. I need them all in a single table,( it doesn't matter if it's in the same database or not) but excluding the duplicate rows.

They all have the same no. of columns and the same column names.

Please help me create a function, because I have tried UNION But it just does not work, I have also tried UNIQUE, and it also does not work.

This is my function that involves UNIQUE:

def create_table():
    c.execute('CREATE TABLE IF NOT EXISTS ring(members INTEGER,distance REAL,UNIQUE(members, distance))')

it always spews out an error. BUt please, do not focus on my sample code too much, I'm looking for better ways.

Thank you so much.

Answer 1

I believe the following will do as you wish (table names would need changing, hopefully explanatory table names used) :-

WITH RECURSIVE all4 AS(
SELECT * FROM 
    ring1
    UNION SELECT * FROM ring2 UNION SELECT * FROM ring3 UNION SELECT * FROM ring4
)
INSERT OR IGNORE INTO newring SELECT * FROM all4 
;
SELECT * FROM newring;

The following demonstrates the above :-

DROP TABLE If EXISTS ring1;
DROP TABLE If EXISTS ring2;
DROP TABLE If EXISTS ring3;
DROP TABLE If EXISTS ring4;
DROP TABLE IF EXISTS newring;
CREATE TABLE IF NOT EXISTS ring1(othercolumn TEXT,members INTEGER,distance REAL);
CREATE TABLE IF NOT EXISTS ring2(othercolumn TEXT,members INTEGER,distance REAL);
CREATE TABLE IF NOT EXISTS ring3(othercolumn TEXT,members INTEGER,distance REAL);
CREATE TABLE IF NOT EXISTS ring4(othercolumn TEXT,members INTEGER,distance REAL);
CREATE TABLE IF NOT EXISTS newring(othercolumn TEXT,members INTEGER,distance REAL, UNIQUE(members,distance));
INSERT INTO ring1 VALUES
    ('A',10,30.5),('B',11,30.6),('C',12,30.7),('D',13,30.8), -- duplicates not the othercolumn though
    ('X',100,15.0),('Y',110,17.0),('Z',120,19.0)
;
INSERT INTO ring2 VALUES
    ('E',10,30.5),('F',11,30.6),('G',12,30.7),('H',13,30.8), -- duplicates not the othercolumn though
    ('X',200,15.0),('Y',210,17.0),('Z',220,19.0)
;
INSERT INTO ring3 VALUES
    ('I',10,30.5),('J',11,30.6),('K',12,30.7),('L',13,30.8), -- duplicates not the othercolumn though
    ('X',300,15.0),('Y',310,17.0),('Z',320,19.0)
;
INSERT INTO ring4 VALUES
    ('M',10,30.5),('N',11,30.6),('O',12,30.7),('P',13,30.8), -- duplicates not the othercolumn though
    ('X',400,15.0),('Y',410,17.0),('Z',420,19.0)
;
WITH RECURSIVE all4 AS(
SELECT * FROM 
    ring1
    UNION SELECT * FROM ring2 UNION SELECT * FROM ring3 UNION SELECT * FROM ring4
)
SELECT * FROM all4 
;
WITH RECURSIVE all4 AS(
SELECT * FROM 
    ring1
    UNION SELECT * FROM ring2 UNION SELECT * FROM ring3 UNION SELECT * FROM ring4
)
INSERT OR IGNORE INTO newring SELECT * FROM all4 
;
SELECT * FROM newring;

The result being :-

That is of the 28 rows 16 were added.

16 as each of the four base tables has 3 unique rows (so 12 rows), each table has 4 rows that have the same member/distance combination.

Thus, only 4 of those 16 will be added due to the UNIQUE constraint and the INSERT OR IGNORE, and therefore 16 rows in total added.

The intermediate query ie the UNION of the 4 tables produces 28 rows as per :-