我正在尝试加入两个数据集
Ds1
+-------------+-----------------+-----------+
| countryName|countryPopulation|countrySize|
+-------------+-----------------+-----------+
| China| 1210004992| 9596960|
| India| 952107712| 3287590|
| UnitedStates| 266476272| 9372610|
| Indonesia| 206611600| 1919440|
| Brazil| 162661216| 8511965|
| Russia| 148178480| 17075200|
| Pakistan| 129275664| 803940|
| Japan| 125449704| 377835|
| Bangladesh| 123062800| 144000|
| Nigeria| 103912488| 923770|
| Mexico| 95772464| 1972550|
| Germany| 83536112| 356910|
| Philippines| 74480848| 300000|
| Vietnam| 73976976| 329560|
| Iran| 66094264| 1648000|
| Egypt| 63575108| 1001450|
| Turkey| 62484480| 780580|
| Thailand| 58851356| 514000|
|UnitedKingdom| 58489976| 244820|
| France| 58317448| 547030|
+-------------+-----------------+-----------+
Ds2:
+------------+-----------------+-----------+
| countryName|countryPopulation|countrySize|
+------------+-----------------+-----------+
| China| 1210004992| 9596960|
| India| 952107712| 3287590|
|UnitedStates| 266476272| 9372610|
| Indonesia| 206611600| 1919440|
| Brazil| 162661216| 8511965|
| Russia| 148178480| 17075200|
| Pakistan| 129275664| 803940|
| Japan| 125449704| 377835|
| Bangladesh| 123062800| 144000|
| Nigeria| 103912488| 923770|
| Germany| 83536112| 356910|
| Vietnam| 73976976| 329560|
| Iran| 66094264| 1648000|
| Thailand| 58851356| 514000|
| France| 58317448| 547030|
| Italy| 57460272| 301230|
| Ethiopia| 57171664| 1127127|
| Ukraine| 50864008| 603700|
| Zaire| 46498540| 2345410|
| Burma| 45975624| 678500|
+------------+-----------------+-----------+
当我执行以下操作时,我将输出
Dataset<Row> ds3 = ds2.filter(ds2.col("countryPopulation").cast("int").$greater(100000))
.join(ds1, ds1.col("countrySize")
.equalTo(ds2.col("countrySize")));
ds3.show();
但是当我执行以下操作时,我得到了错误
Dataset<Row> ds3 = ds2.filter(ds2.col("countryPopulation").cast("int").$greater(100000))
.join(ds1, ds1.col("countrySize").cast(DataTypes.IntegerType)
.equalTo(ds2.col("countrySize").cast(DataTypes.IntegerType)), "inner");
ds3.show();
错误:
Exception in thread "main" org.apache.spark.sql.AnalysisException: Detected implicit cartesian product for INNER join between logical plans
Project [country#6.name AS countryName#2, country#6.population AS countryPopulation#3, country#6.area AS countrySize#4]
+- Filter (isnotnull(country#6) && (Contains(country#6.name, a) && ((cast(country#6.population as int) > 100000) && (cast(country#6.area as int) = cast(country#6.area as int)))))
+- Generate explode(countries#0.country), [0], false, t, [country#6]
+- Relation[countries#0] json
我能知道吗,请问应该如何同时投射和加入..?以及为什么会收到此错误??
此“逻辑计划之间用于内部联接的检测到的隐式笛卡尔积”的错误含义是什么?
答案 0 :(得分:0)
我已经看到,当where条件包含带有来自两个数据帧的参数的函数调用时,就会发生笛卡尔联接。类似于df1.join(df2, aFunction(df1.column, df2.column)
。在这里我看不到确切的位置,但我怀疑这种情况正在发生。
请尝试以下操作,以将功能应用于选择条件而不是位置。
Dataset<Row> ds1_1 = ds1.select(col("countrySize").cast(DataTypes.IntegerType).as("countrySize")) // add all columns here
Dataset<Row> ds2_1 = ds2.select(col("countrySize").cast(DataTypes.IntegerType).as("countrySize"),ds2.col("countryPopulation").cast("int").as("countryPopulation")) // add all columns here
Dataset<Row> ds3 = ds2.filter(ds2_1.col("countryPopulation").$greater(100000))
.join(ds1_1, ds1_1.col("countrySize")
.equalTo(ds2_1.col("countrySize")), "inner");
ds3.show();