topic Re: using the VA data preparation tool (data query) to manipulate lasr tables in SAS Visual Analytics

using the VA data preparation tool (data query) to manipulate lasr tables

mhamlett — Sun, 05 Mar 2017 16:12:38 GMT

my question concerns the amount of memory the results of a data query occupies in lasr as opposed to the tables being loaded to lasr.

example

i have 2 tables in lasr (customer information, sales detail data by day)

for example sake, lets say the customer is 500 meg in lasr and the sales detail is 500 meg in memory

each of the 2 tables stands alone in lasr and occupies in memory space in lasr ( 1 gig)

if i join these tables using the "Data Preparation Tool" to create a 3rd table of sales by customer, does the 3rd table occupy more than the physical memory space occupied by the individual tables? (at least another gig of memory)

or is this type of join more "logical" whereas this is a virtual join and does not consume lasr memory in proportion to the physical tables.

thx

Re: using the VA data preparation tool (data query) to manipulate lasr tables

SASKiwi — Mon, 06 Mar 2017 05:52:44 GMT

I'm guessing you are asking if it is possible to store an SQL query VIEW in a LASR library. The view takes up virtually no memory but would dynamically join the two input tables when used. This reference confirms that views are possible:

http://support.sas.com/documentation/cdl/en/vaug/68648/HTML/default/viewer.htm#p0ci6glvfbc1kin1ipn3uae2xnyv.htm

I'm not sure what the impact on performance would be using a view - it would be slower than a table. Try it and let us know how you get on.

Re: using the VA data preparation tool (data query) to manipulate lasr tables

LinusH — Mon, 06 Mar 2017 10:23:22 GMT

Another option could be to use Star Schemas:

http://support.sas.com/documentation/cdl/en/vaug/68648/HTML/default/viewer.htm#n1xnkytqn4c3kyn1qfco4krqkl9f.htm