Question

FILE COntent（test.txt）：

Some    specific    column      value: x192.168.1.2     blah       blah
Some    specific    row        value: y192.168.1.3      blah       blah
Some    specific    field      value: z192.168.1.4     blah      blah

猪查询：

A = LOAD 'test.txt' USING PigStorage('\t') AS (data1: chararray , data2: chararray , data3: chararray, data4: chararray , data5: chararray , data6: chararray);

B = foreach A generate data3, data4;

C = filter B by data3 matches 'row';

D = foreach C generate data4;

E = foreach D generate TOKENIZE(data4);

输出：

((value:), (y192.168.1.3))

现在我想在这个输出包中提取特定的元组，比如第二元组（y192.168.1.3）。在此之后我想提取IP地址。我正在尝试使用UDF，但卡住了。

Answer 1

您可以使用Flatten Operator展平行李，然后使用过滤器提取IP地址。

E = foreach C generate flatten(TOKENIZE(data4));
F = filter E by $0 matches '.\\d+\\.\\d+\\.\\d+\\.\\d+'

希望这有帮助

Answer 2

这就是我要做的事。

PIG脚本

A = LOAD 'test.txt' USING PigStorage('\t') AS (data1: chararray , data2: chararray , data3: chararray, data4: chararray , data5: chararray , data6: chararray);
B = foreach A generate data3, data4;
C = filter B by data3 matches 'row';
D = foreach C generate data4;
E = foreach D generate REGEX_EXTRACT($0,'value: .([0-9]+\\.[0-9]+\\.[0-9]+\\.[0-9]+).*', 1);

输出

(192.168.1.3)

如果需要，您可以使用更疯狂的正则表达式来提取IP地址：Extract ip addresses from Strings using regex

Answer 3

public class someClass extends EvalFunc<String>
{
   public String exec(Tuple input) throws IOException {
     DataBag bag = (DataBag)input.get(0);
     Iterator<Tuple> it = bag.iterator();
     Tuple tup;
     for(int i = 0; i < 2; i++)
     {
       tup = it.next();
     }
     String ipString = tup.get(0);
     String ip = //get ip from string with a regex
     return ip;
   }
 }

当然你应该添加一些输入检查（空输入，包大小为1等）并保护代码。

从PIG中提取袋中的元组

3 个答案: