tricks I fall into in exercises ：
4.1 It's better for a relation to have a primary key to avoid duplicates ...
4.11 with declaration : with with_name(attri1, attri2, ..., attrin) as (...here fill query results...)
4.13 answer is very simple, append depth as the last && new attribute is ok .
technic 1 to express 'for all X, P(X)' ->> 'not exist X such that not P(X)'.
Q1 difference between SQL && Embedded SQL(in C)?
1.use a declare cursor , result is not computed
2.use open && fetch to obtain result.
Q2 principle(or process) of fetching query result using Embedded SQL (in C)?
open -> while(have item) -> fetch -> next ->go to while.
not familiar:
Dynamic SQL, ODBC, JDBC, Metadata
sth. funny:
That is, we wish to find employees who are supervised directly by Jones, or who are supervised by someone
who is supervsed by Jones, or who are supervised by someone who is supervised by someone who is supervised Jones, and so on.
outline:
additional built-in data types: date , time, timestamp
user-defined types : create type ... as exist type final
lob
catalog.schema.relation
integrity constraint
constraint : not null, unique...(candicate key), check(predicate)