A query identifier can be declared using the -ref=... construct. The query identifier itself must be a string that starts with ’$’. Declaring a query identifier has two consequences:
- The query is not executed
- The query can be employed by another query or command
Queries are saved on a persistent ’query stack’ on the server side.
A usual application of query identifiers is to perform logical operations on sets. For example, execute sequentially the following commands;
query_reset # cleaning up the query stack (return nothing)
seq_query seq_name=sw:VAV_HUMAN,sw:VAV_RAT,sw:VAV_MOUSE -ref=$X # $X: three VAV oncogenes
seq_query seq_name=sw:VAV_HUMAN,sw:VAV_RAT -ref=$Y # $Y: only VAV from human and rat
seq_query seq_name=VAV_HUMAN -ref=$Z # $Z: only VAV from human
seq_query seq_name=$X and_seq_name=$Y not_seq_name=$Z # $X AND $Y NOT $Z
Here below is the same example in one block of text. You can cut and paste it to your command line client. Or you can import it into the HTML interface using the ad-hoc button. The lines that starts with ’//’ are sticky comments that bind to the next query definition to endowed the HTML interface with human-readable comments.