Question

这是我的观点，它将3个表中的列放在一起

        db.execSQL("CREATE VIEW " + viewComps +
            " AS SELECT " + COMPANY + "." + colCompID + " AS _id," +
            " " + ACCOUNTS + "." + colName + "," +
            " " + COMPANY + "." + colCompClass + "," +
            " " + PAYMENTS + "." + colGroupID + "," +
            " " + PAYMENTS + "." + colPayDue + "," +
            " " + PAYMENTS + "." + colDateDue + "" +
            " FROM " + PAYMENTS + ", " + COMPANY +
            " JOIN " + ACCOUNTS + " ON " + PAYMENTS + "." + colGroupID + " = " + ACCOUNTS + "." + colID );

问题

我有3家公司A，B和C
Acc 1分配给A，1 Acc = 1 Company ONLY
插入Acc 1并分配给A，问题是即使我将其分配给A，它也会被复制为B和C.

结果：

  Acc 1 | Company A | Payment | Date
  Acc 1 | Company B | Payment | Date
  Acc 1 | Company C | Payment | Date

应该是什么：

  Acc 1 | Company A | Payment | Date

我插入到数据库中的每个其他帐户都会执行相同操作，并导致我的公司包含所有帐户的副本，而不管该帐户分配给哪个公司。

问题

我的视图有什么问题？我不明白为什么它会复制所有条目，将每个帐户的副本放在每个公司中。有人能告诉我我的错误在哪里吗？我对此很新，可以在这方面使用一些指针。我非常有信心这只是这个观点的一个问题，因为我有另一个基于公司显示帐户的活动，它至少在那里工作得很好。

如果需要，我会在下面的3个表中保留VIEW参考：

    db.execSQL("CREATE TABLE " + COMPANY + " (" + colCompID + " INTEGER PRIMARY KEY AUTOINCREMENT, " +
            colCompClass + " TEXT)");

    db.execSQL("CREATE TABLE " + ACCOUNTS + " (" + colID + " INTEGER PRIMARY KEY AUTOINCREMENT, " +
            colName + " TEXT, " +
            colComp + " INTEGER NOT NULL," +
            colAmount + " INTEGER, " +
            colPurpose + " TEXT, " +
            colTerms + " INTEGER NOT NULL, " +
            colPeriod + " INTEGER NOT NULL, " +
            colBalance + " INTEGER, "+
            colStatus + " INTEGER NOT NULL," +
            colDate + " TEXT, " +
            colEditDate + " TEXT, " +
            colRemarks + " TEXT, " +
            "FOREIGN KEY (" + colComp + ") REFERENCES " + COMPANY + " (" + colCompID + " )" + "ON DELETE CASCADE," +
            "FOREIGN KEY (" + colTerms + ") REFERENCES " + TERMS + " (" + colTermsID + " )" + "ON DELETE CASCADE," +
            "FOREIGN KEY (" + colPeriod + ") REFERENCES " + PERIODS + " (" + colPeriodID + ") " + "ON DELETE CASCADE,"+
            "FOREIGN KEY (" + colStatus + ") REFERENCES " + STATUS + " (" + colStatusID + ") ON DELETE CASCADE);");

    db.execSQL("CREATE TABLE " + PAYMENTS + " (" + colPayID + " INTEGER PRIMARY KEY AUTOINCREMENT, " +
            colGroupID + " INTEGER NOT NULL, " +
            colPayBal + " TEXT, " +
            colInterest + " TEXT, " +
            colPayDue + " TEXT, " +
            colDateDue + " TEXT, " +
            colPayDate + " TEXT, " +
            "FOREIGN KEY (" + colGroupID + ") REFERENCES " + ACCOUNTS + " (" + colID + ") ON DELETE CASCADE);");

Answer 1

当您执行没有约束的连接时，它会在两个数据集之间生成笛卡尔积 - 左侧的每一行与右侧的每一行组合，因此两个表的大小分别为M和N join将产生大小（M x N）的结果。

您的视图查询执行两个连接。一个连接（ACCOUNTS）具有限制结果集大小的约束，但另一个（COMPANY）没有。这意味着当您加入PAYMENTS和COMPANY时，您将获得支付和公司的每种组合的实例。假设每个中有3个，它产生了这个：

Company A | Payment 1
Company A | Payment 2
Company A | Payment 3
Company B | Payment 1
Company B | Payment 2
Company B | Payment 3
Company C | Payment 1
Company C | Payment 2
Company C | Payment 3

在我看来，您认为自己获得重复帐户的观察实际上是您有重复付款的观察结果。这些帐户是根据付款加入的，因此这是第一次产生错误结果的加入。

您需要的是两个连接的连接约束。看看你的架构，看起来付款和帐户有关系，所以我认为正确的查询看起来像这样：

CREATE VIEW viewComps AS
SELECT ACCOUNTS.colName, COMPANY.colCompId AS _id, COMPANY.colCompClass, PAYMENTS.colGroupID, PAYMENTS.colPayDue, PAYMENTS.colDateDue
FROM ACCOUNTS
JOIN COMPANY ON (COMPANY.colCompId = ACCOUNTS.colComp)
JOIN PAYMENTS ON (PAYMENTS.colGroupID = ACCOUNTS.colID);

CREATE VIEW中的SQLite元素重复

1 个答案: