Oracle float (B)

From: http://www.cnblogs.com/boulder19830907/archive/2008/01/31/1059627.html

Someone asked me today, after designing the field type to float (2) and inserting 93.5 data, why is it 90?

In order to illustrate this problem, let's look at a paragraph (http://www.cnoug.org/viewthread.php? Tid = 56643): Oracle online help says: Float (B) specifies a floating-point number with binary precision B. The precision
B can range from 1 to 126. to convert from binary to decimal precision, multiply B by 0.30103.

Based on this section, we can see that 2 in float (2) is a binary precision, rather than the commonly used decimal precision. The Conversion Relationship between them is: Binary precision = int (B * 0.30103), so the actual precision here should be equal to int (2*0.30103) = 0, that is, the precision after the decimal point is 0.

Back to our original problem, 93.5 is converted to floating point 9.35*10 ^ 1. At this time, after 9.35 decimal points, the precision is 0 to 9, and then it is changed to 9*10 ^ 1 = 90.

Similarly, we can design it as float (10), so int (10*0.30103) = 3. Therefore, if we insert 93.5, we will get 9.35 -- the precision is 3 --> 9.350, 9.350*10 ^ 1 = 93.5. If 13884.2 is inserted, 1.38842 -- the precision is 3 --> 1.388, 1.388*10 ^ 4 = 13880.

We can see this conversion rule more clearly from the following tests.

Connected to Oracle9i Enterprise Edition Release 9.0.1.1.1
Connected as aspire

SQL>
SQL> Create Table hjm_float_test
2 (a float (2 ),
3 B float (10 ),
4 C float,
5 d number );

Table created

SQL> insert into hjm_float_test (A, B, C, D) values (93.5, 93.5, 93.5, 93.5 );

1 row inserted

SQL> insert into hjm_float_test (A, B, C, D) values (93.665, 93.665, 93.665, 93.665 );

1 row inserted

SQL> insert into hjm_float_test (A, B, C, D) values (96.5, 96.5, 96.5, 96.5 );

1 row inserted

SQL> insert into hjm_float_test (A, B, C, D) values (13884.2, 13884.2, 13884.2, 13884.2 );

1 row inserted

SQL> commit;

Commit complete

SQL> select * From hjm_float_test;

A B c d
----------------------------------------------------------------------------------------------------------
90 93.5 93.5 93.5
90 93.67 93.665 93.665
100 96.5 96.5 96.5
1000 13880 13884.2 13884.2

SQL> DESC hjm_float_test;
Name type nullable default comments
---------------------------------
A float y
B float y
C float y
D Number y

SQL>

Please note that, although I designed the tables as float (2) and float (10), when I was in the DESC table (the last 7 columns of the Code) you cannot see such a design structure. Special attention should be paid to such problems !!

Finally, let's review,Remember that B in float (B) is the binary precision, and its conversion rule is binary precision = int (B * 0.30103 ).