2009/5/17 João Carlos dos Reis Baptista <joao.bapti...@gravames.com.br>:
> I have never used PIPE but I Will read about it, you may bet, but if there is > another simple way, please drop me some lines about it. Although it looks like you just need something like Fran shows to catalog the file again, here's some plumbing to approach the data from the VM side. Oh, and you need the CMS Pipelines Runtime Library from http://vm.marist.edu/~pipeline for this. When you have your VSE volume linked R/O in a VM user at address 200, the following pipeline gets you the blocks. The 'nn' is the starting cylinder of the extent, and 'mm' is the number of tracks to read. After the deblocking, those two stages get rid of R0 pipe trackread 200 nn 0 mm | trackdeblock | locate 8 | nlocate 5.1 x00 | .. The remainder of the pipeline should now split the records in count area, key and data. And if your data records are 4K it is easy to keep them .. ckddeblock | locate 4096 | ... The rest depends on the format of the data blocks. In most cases there will be a pipeline deblocking stage that can pull the actual data out of the records. But if you can just catalog the dataset again on VSE, that's probably much easier... Rob