[英]Compare a list values with case class using Scala and Spark
I have a dataframe like below.我有一个如下所示的数据框。
+-------+------+-------+-------+
| num1 | num2 | x | y |
+-------+------+-------+-------+
| 25 | 10 | a&c | i&j&k |
| 35 | 15 | a&b&d | i&k |
+-------+------+-------+-------+
I have another data frame structure with the headers like,我有另一个带有标题的数据帧结构,例如,
num1, num2, a, b, c, d, i, j, k
I want to split the column data of x and y from the symbol "&".我想从符号“&”中拆分 x 和 y 的列数据。 Then check whether the split data are matching with the headers above, also considering the columns num1 and num2.
然后检查拆分数据是否与上面的标题匹配,还要考虑列 num1 和 num2。 If it so fill the values with 1 else with 0.
如果是这样,则用 1 填充值,否则用 0 填充。
The required output is:所需的输出是:
+-------+------+---+---+---+---+---+---+---+
| num1 | num2 | a | b | c | d | i | j | k |
+-------+------+---+---+---+---+---+---+---+
| 25 | 10 | 1 | 0 | 1 | 0 | 1 | 1 | 1 |
| 35 | 15 | 1 | 1 | 0 | 1 | 1 | 0 | 1 |
+-------+------+---+---+---+---+---+---+---+
I have achieved the above output in a method like following.我已经通过以下方法实现了上述输出。 I created another data frame same like the first data frame but the x and y contains with an array of split data like following.
我创建了另一个与第一个数据帧相同的数据帧,但 x 和 y 包含一个拆分数据数组,如下所示。
+------+-------+---------+---------+
| num1 | num2 | x | y |
+------+-------+---------+---------+
| 25 | 10 | [a,c] | [i,j,k] |
| 35 | 15 | [a,b,d] | [i,k] |
+------+-------+---------+---------+
Then followed the solution in this question然后按照这个问题中的解决方案
Although it gives me the exact solution, it is ineffective when it comes to the case where there are lot of columns like x and y.虽然它给了我确切的解决方案,但当涉及到像 x 和 y 这样的列很多的情况时,它是无效的。
So now I want to create a case class and match the header values with the data in x,y columns by splitting them to a list.所以现在我想创建一个案例类,并通过将它们拆分为一个列表来将标题值与 x,y 列中的数据进行匹配。 Is it possible or is there any other solution?
是否有可能或有其他解决方案? Can someone help me?
有人能帮我吗?
After trying several methods at last I came up with the following solution.最后尝试了几种方法后,我想出了以下解决方案。 I found my solution by adding some few changes to the answer for this question: Compare rows of an array column with the headers of another data frame using Scala and Spark .
我通过对这个问题的答案添加一些更改找到了我的解决方案: 使用 Scala 和 Spark 将数组列的行与另一个数据帧的标题进行比较。 It worked for multiple array columns also.
它也适用于多个数组列。 This is the code for it.
这是它的代码。
val df = Seq((25, 10, "a&c", "i&j&k"), (35, 15, "a&b&d", "i&k")
.toDF("num1", "num2", "x", "y")
val dfProcessed = df.withColumn("x", split($"x", "&"))
.withColumn("y", split($"y", "&"))
.select("num1", "num2", "x", "y")
val headers = Seq("a", "b", "c", "d", "i", "j", "k")
val report = dfProcessed.select(Seq("num1", "num2").map(col) ++ headers.map(line => array_contains('x, line)
|| array_contains('y, line) as line) : _*)
report.show()
I think this may help you.我想这可能对你有帮助。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.