Queries

``Queries'' are used to extract and combine information from existing tables to produce new tables.

Relational databases provide these ``operators'' that can be composed into queries:

• projection -- discard some columns
• selection -- discard some rows
• cartprod -- cartesian product of two tables
• renaming -- change the name of a column
• set difference
• set union

Example query:

{ (#accn:   x.#accn,
   #uid:    x.#uid, 
   #title:  x.#title, 
   #seq:    x.#seq,
   #f-start: y.#start, 
   #f-end  : y.#end, 
   #f-name : y.#name, 
   #f-anno : y.#anno)
| \x <- genpepts, \y <- features,
  y.#uid = x.#uid }

Can you see what operators made up the above query?

Note that a query is a sequence of purely structural operations. It does not care what the data is, what a column means. It merely, ``select'', ``project'', etc.