【问题标题】:replace comma(,) only if its inside quotes("") in Pig仅当它在 Pig 中的内部引号(“”)时才替换逗号(,)
【发布时间】:2017-07-20 09:10:31
【问题描述】:

我有这样的数据:

1,234,"john, lee", john@xyz.com

我想使用 pig 脚本删除 "" 内的空格。这样我的数据将如下所示:

1,234,john lee, john@xyz.com

我尝试使用 CSVExcelStorage 加载此数据,但我还需要使用 CSVExcelStorage 不支持的“-tagFile”选项。所以我打算只使用 PigStorage,然后替换引号内的任何逗号 (,)。 我坚持这一点。非常感谢任何帮助。谢谢

【问题讨论】:

    标签: apache-pig


    【解决方案1】:

    以下命令会有所帮助:

    csvFile = load '/path/to/file' using PigStorage(',');
    result = foreach csvFile generate $0 as (field1:chararray),$1 as (field2:chararray),CONCAT(REPLACE($2, '\\"', '') , REPLACE($3, '\\"', '')) as field3,$4 as (field4:chararray);
    

    输出:

    (1,234,john lee, john@xyz.com)

    【讨论】:

    【解决方案2】:

    将其加载到单个字段中,然后使用 STRSPLIT 和 REPLACE

    A = LOAD 'data.csv' USING TextLoader() AS (line:chararray);
    B = FOREACH A GENERATE STRSPLIT(line,'\\"',3); 
    C = FOREACH B GENERATE REPLACE($1,',','');
    D = FOREACH C GENERATE CONCAT(CONCAT($0,$1),$2); -- You can further use STRSPLIT to get individual fields or just CONCAT
    E = FOREACH D GENERATE STRSPLIT(D.$0,',',4);
    DUMP E;
    

    一个

    1,234,"john, lee", john@xyz.com
    

    B

    (1,234,)(john, lee)(, john@xyz.com)
    

    C

    (1,234,)(john lee)(, john@xyz.com)
    

    D

    (1,234,john lee, john@xyz.com)
    

    E

    (1),(234),(john lee),(john@xyz.com)
    

    【讨论】:

      【解决方案3】:

      我找到了完美的方法来做到这一点。一个非常通用的解决方案如下:

      data = LOAD 'data.csv' using PigStorage(',','-tagFile') AS (filename:chararray, record:chararray);
      
      /*replace comma(,) if it appears in column content*/
      replaceComma = FOREACH data GENERATE filename, REPLACE (record, ',(?!(([^\\"]*\\"){2})*[^\\"]*$)', '');
      
      /*replace the quotes("") which is present around the column if it have comma(,) as its a csv file feature*/
      replaceQuotes = FOREACH replaceComma GENERATE filename, REPLACE ($4,'"','') as record;
      

      详细用例可在my blog获取

      【讨论】:

        猜你喜欢
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        相关资源
        最近更新 更多