This post was tested on an Oracle 9 database. First create a small table and insert a few rows:
SQL> create table emp2 (name varchar2(10),
3 dob date,
4 salary number)
5 /
Table created.
SQL> insert into emp values
2 ('Andrew','01-JAN-1960',10000)
3 /
1 row created.
SQL> insert into emp values
2 ('Brian','02-FEB-1965',20000)
3 /
1 row created.
SQL> insert into emp values
2 ('Colin','03-MAR-1970',30000)
3 /
1 row created.
SQL> insert into emp values
2 ('David','04-APR-1975',40000)
3 /
1 row created.
SQL> insert into emp values
2 ('Edward','05-MAY-1980',50000)
3 /
1 row created.
SQL>
You can count the number of rows in a table as follows:
SQL> select count(*) from emp
2 /
COUNT(*)
----------
5
SQL>
This is how you select all the columns in every row:
SQL> select * from emp
2 /
NAME DOB SALARY
---------- --------- ----------
Andrew 01-JAN-60 10000
Brian 02-FEB-65 20000
Colin 03-MAR-70 30000
David 04-APR-75 40000
Edward 05-MAY-80 50000
SQL>
The next query shows some of the columns from all of the rows:
SQL> select name, dob from emp
2 /
NAME DOB
---------- ---------
Andrew 01-JAN-60
Brian 02-FEB-65
Colin 03-MAR-70
David 04-APR-75
Edward 05-MAY-80
Finally, to see all of the columns from some of the rows, you need to use a where clause:
SQL> select * from emp where name = 'Andrew'
2 /
NAME DOB SALARY
---------- --------- ----------
Andrew 01-JAN-60 10000
SQL>
No comments:
Post a Comment