I am adding the user mailing list so more people can benefit from it. I have not done what you have asked myself but in theory it should be possible with overriding the file name format, although it is not the cleanest way to do it, so I would suggest you file a JIRA for us to clean it up.
FileNameFormat https://github.com/apache/storm/blob/master/external/storm-hdfs/src/main/java/org/apache/storm/hdfs/bolt/format/FileNameFormat.java has two methods. getPath gives the base path everything is placed in, and getName takes the rotation and timestamp to come up with the name for a file that things are written to. Luckily Hadoop is happy to treat a new Path("base", "a/b") and a file named "b" in a directory named "base/a". So you can simply override getName and use the timestamp to create something in the format of timestamp + "/" + whatever else you want. - Bobby On Wednesday, May 13, 2015 8:10 PM, "zhangl...@douyu.tv" <zhangl...@1st.norecord.glb0.lxdns.com> wrote: #yiv6661092873 body {line-height:1.5;}#yiv6661092873 body {font-size:10.5pt;color:rgb(0, 0, 0);line-height:1.5;}Dear bobby: I am confused when I use strom-hdfs. I want to put the data stored in different folders for different dates. But I do not know how to configure. Can you give me an example? Thanks. zhangl...@douyu.tv