INSERT ... 冲突时什么都不做 - 读取 csv 并生成外键表

Question

我正在尝试读取包含艺术家、专辑、歌曲和标签列的 csv 文件。

我希望像这样填充artist_album_song 表：

|artist_id|album_id|song_id|
|---------|--------|-------|
|   1     |     1  |     1 |
|   1     |     1  |     2 |
|   1     |     2  |     1 |
...
|  12     |     1  |     1 |
...

我已经设计并正在尝试填充以下表格。当我在 csv 中读取时，问题是填充 artist_album_song 表中的外键。

在下面使用的 INSERT 语句中插入此表的最佳方法是什么（返回语法错误）？谢谢。

create table artists (
    artist_id SERIAL PRIMARY KEY,
    artist VARCHAR(100) NOT NULL UNIQUE
);

create table albums (
    album_id SERIAL PRIMARY KEY,
    album VARCHAR(100) NOT NULL UNIQUE
);

create table songs (
    song_id SERIAL PRIMARY KEY,
    song VARCHAR(250) NOT NULL UNIQUE
);

create table tags (
    tag_id SERIAL PRIMARY KEY,
    tag VARCHAR(100) NOT NULL UNIQUE
);

create table artists_albums_songs (
    artist_id INTEGER NOT NULL,
    album_id INTEGER NOT NULL,
    song_id INTEGER NOT NULL,
    FOREIGN KEY (artist_id) REFERENCES artists(artist_id),
    FOREIGN KEY (album_id) REFERENCES albums(album_id),
    FOREIGN KEY (song_id) REFERENCES songs(song_id),
    PRIMARY KEY (artist_id, album_id, song_id)
);

create table songs_tags (
    song_id INTEGER NOT NULL,
    tag_id INTEGER NOT NULL,
    FOREIGN KEY (song_id) REFERENCES songs(song_id),
    FOREIGN KEY (tag_id) REFERENCES tags(tag_id),
    PRIMARY KEY (song_id, tag_id)
);

在尝试了以下链接中的各种语句变体后，我仍然无法使其正常工作。

我已经尝试了以下语句，但我不断收到错误消息。第一个返回错误：

org.postgresql.util.PSQLException: ERROR: syntax error at or near "ON" Position: 161;

161是指下面SQL语句中的第161个字符吗？

INSERT INTO artists_albums_songs
SELECT artist_id, album_id, song_id 
FROM artists a 
    JOIN albums b
        ON a.artist = ?
        AND b.album = ?
    JOIN songs c
        ON c.song = ?
    ON DUPLICATE (artist_id, album_id, song_id) DO NOTHING;

INSERT INTO artists_albums_songs
SELECT artist_id, album_id, song_id 
FROM artists a 
    JOIN albums b
        ON a.artist = ?
        AND b.album = ?
    JOIN songs c
        ON c.song = ?
    WHERE NOT EXISTS (
        SELECT * 
        FROM artists_albums_songs
        WHERE * = ?, ?, ?)

INSERT INTO artists_albums_songs
SELECT artist_id, album_id, song_id 
FROM artists a 
    JOIN albums b
        ON a.artist = ?
        AND b.album = ?
    JOIN songs c
        ON c.song = ?
    ON CONFLICT (song_id) IGNORE;

编辑：如果我删除上面 3 个 INSERT 语句的最后一行，它可以工作，但是当它遇到重复时，它会说：

org.postgresql.util.PSQLException: ERROR: duplicate key value violates unique constraint "artists_albums_songs_pkey"
  Detail: Key (artist_id, album_id, song_id)=(1, 1, 1) already exists.

Insert, on duplicate update in PostgreSQL?

Use INSERT ... ON CONFLICT DO NOTHING RETURNING failed rows

How to UPSERT (MERGE, INSERT ... ON DUPLICATE UPDATE) in PostgreSQL?

Answer 1

编辑 1： 我刚刚意识到我可以用 Java 处理这些错误！所以我的解决方案只是添加一个catch 语句来处理Duplicate SQLException

private <T> void insertIntoArtistAlbumSong(T artist, T album, T song) throws SQLException {

    try {

        String artString = artist.toString();
        String albString = album.toString();
        String songString = song.toString();

        // Create SQL insert statement
        String stm =
                "INSERT INTO artists_albums_songs " +
                        "SELECT artist_id, album_id, song_id " +
                        "FROM artists a " +
                        "JOIN albums b " +
                        "ON a.artist = ? " +
                        "AND b.album = ? " +
                        "JOIN songs c " +
                        "ON c.song = ? ;";


        PreparedStatement pstmt = connection.prepareStatement(stm);

        // Set values in prepared statement
        pstmt.setString(1, artString);
        pstmt.setString(2, albString);
        pstmt.setString(3, songString);

        // Insert into table
        pstmt.executeUpdate();

    // ADDED THIS CATCH STATEMENT!
    } catch (SQLException e){
        System.out.println(e.getSQLState());
    }
}

---

好的，所以我找到了一个解决方案，但它只适用于填充表格（这是我实际上必须做的）。

1. 删除原来的artists_albums_songs[1]表

2. 新建artists_albums_songs[2]表没有约束：

   CREATE TABLE artists_albums_songs (
       artist_id INTEGER NOT NULL,
       album_id INTEGER NOT NULL,
       song_id INTEGER NOT NULL
   );
   

3. 然后我用以下语句（通过 JDBC）填充 new 表[2]：

   INSERT INTO artists_albums_songs
   SELECT artist_id, album_id, song_id 
   FROM artists a 
       JOIN albums b
           ON a.artist = ?
           AND b.album = ?
       JOIN songs c
           ON c.song = ?;
   

4. 创建一个tmp[3] 表带有约束（通过 psql 命令行）：

   CREATE TABLE tmp (
       artist_id INTEGER NOT NULL,
       album_id INTEGER NOT NULL,
       song_id INTEGER NOT NULL,
       FOREIGN KEY (artist_id) REFERENCES artists(artist_id),
       FOREIGN KEY (album_id) REFERENCES albums(album_id),
       FOREIGN KEY (song_id) REFERENCES songs(song_id),
       PRIMARY KEY (artist_id, album_id, song_id)
   );
   

5. 仅将新artists_albums_songs[2] 中的不同 行插入tmp[3]（通过psql）：

   INSERT INTO tmp SELECT DISTINCT * FROM artists_albums_songs
   ORDER BY artist_id, album_id, song_id ASC;
   

6. 删除新的artists_albums_songs[2] 并将tmp[3] 重命名为artists_albums_songs（通过psql）：

   DROP TABLE artists_albums_songs;
   ALTER TABLE tmp RENAME TO artists_albums_songs;
   

Answer 2

此行有错误：

 ON DUPLICATE (artist_id, album_id, song_id) DO NOTHING;

Postgtresql 使用ON CONFLICT 关键字 https://www.postgresql.org/docs/current/static/sql-insert.html