DSXchange

i have a requirement--- In my first job, after extracting and transforming i will create a dataset which will hold some millions of records. in the second job i will extract the data from previously created dataset and do some transformations and again create a dataset. now the requirement when i pulled the data from the datasest in second job i have to delete the data in the dataset created in first job. this is mainly because of space requirements. I read the user guide and it says we can only delete the datasets from dataset management utility. if this is the case how can i delete the datasets.

From unix "orchadmin rm {dataset}" will do it.

ArndW wrote:From unix "orchadmin rm {dataset}" will do it.

But in the document it is said that using unix rm command only descriptor file of dataset is removed leaving the original huge data file on the memory. is orchadmin rm datset different from rm dataset. if i use orchadmin does it remove all the datafile and descriptor file

That was not the UNIX rm command that Arnd suggested.

It was orchadmin rm controlfile.ds that was suggested, to be executed from a UNIX shell with the appropriate environment variables (such as APT_CONFIG_FILE and APT_ORCHHOME) set to correct values.

ArndW wrote:From unix "orchadmin rm {dataset}" will do it.

But in the document it is said that using unix rm command only descriptor file of dataset is removed leaving the original huge data file on the memory. is orchadmin rm datset different from rm dataset. if i use orchadmin does it remove all the datafile and descriptor file

ray.wurlod wrote:That was not the UNIX rm command that Arnd suggested.

It was orchadmin rm controlfile.ds that was suggested, to be executed from a UNIX shell with the appropriate environment variables (such as APT_CONFIG_FILE and APT_ORCHHOME) set to correct values.

Lets suppose if the dataset name is xxx.ds and was stored in /etl/store.
now where i can find the control file and what are the values i should set for apt_config_file and apt_orchhome. i guess i ca execute the script from
before/after subroutine of stage.

The orchadmin command is used to access and manipulate datasets. One of the options is rm which removes the descriptor and component files. Read the documentation to confirm this. Or perhaps you could actually just try it and see if it works.

Do not forget to set-up the environment variable APT_CONFIG_FILE before running the orchadmin.

mdan wrote:Do not forget to set-up the environment variable APT_CONFIG_FILE before running the orchadmin.

when you say set_up the environment variable.do you mean that it is to be added at the job level...can you please be more clear on seting up the environment variable
thanks

Anand - have you tried it? If you don't have the environment variable set you will get an error message.

just4u_sharath wrote:

ArndW wrote:From unix "orchadmin rm {dataset}" will do it.

what is this control file. is it the same as dataset name. setting apt_config_file and pat_orchhome means what. with what values i have to set. please explain

In your case control file and xxx.ds are synonymous

ArndW wrote:In your case control file and xxx.ds are synonymous

Thankyou
And in my case already apt_config_file adn apt_orchome has a default pathname. so i think i just have to add those two paremeters to my job and using before/after subroutine write the unix command. Am i right.

The UNIX command inherits the ENV settings of the calling job, so you will already have a setting for the APT_CONFIG_FILE.