Why not process data on-the-fly? I guess reading and writing data in HDFS would incur high latency.
2014-10-08 15:20 GMT+08:00 snv <[email protected]>: > using mr job to "ETL" the temporary data store in hdfs > > ------------------ > --------------------------------------------- > BestWishes！ > Blog:http://snv.iteye.com/ > > > > > ------------------ 原始邮件 ------------------ > *发件人:* "Dattatrya Moin";<[email protected]>; > *发送时间:* 2014年10月8日(星期三) 下午3:44 > *收件人:* "user"<[email protected]>; > *主题:* Re: Fwd: ETL using Hadoop > > Hi , > > We are extracting data from Oracle database, we are processing that data > (Transformation) i.e. Substring ,Replace and loading modified data back to > Oracle database. > > On Wed, Oct 8, 2014 at 12:13 PM, snv <[email protected]> wrote: > >> where to "E" where to "L"? what do you do in "E" and "T"? >> >> ------------------ >> --------------------------------------------- >> BestWishes！ >> Blog:http://snv.iteye.com/ >> >> >> >> >> ------------------ 原始邮件 ------------------ >> *发件人:* "Dattatrya Moin";<[email protected]>; >> *发送时间:* 2014年10月8日(星期三) 下午3:01 >> *收件人:* "user"<[email protected]>; >> *主题:* Fwd: ETL using Hadoop >> >> >> Hi , >> >> We have our own ETL , but we are planning to use Hadoop for data >> processing as it gives better scalability and performance. As i am new to >> hadoop kindly guide to start with Hadoop. Can we replace ETL using Hadoop . >> And we have any API to connect to Hadoop using .NET. >> >> >> Thanks, >> Dattatrya Moin >> >> >
