ORACLE 中NUMBER類型默認的精度和Scale問題

在ORACLE數據庫中，NUMBER(P,S)是最多見的數字類型，能夠存放數據範圍爲10^-130~10^126（不包含此值)，須要1~22字節(BYTE)不等的存儲空間。P 是Precison的英文縮寫，即精度縮寫，表示有效數字的位數，最多不能超過38個有效數字。S是Scale的英文縮寫，表示從小數點到最低有效數字的位數，它爲負數時，表示從最大有效數字到小數點的位數。有時候，咱們在建立表的時候，NUMBER每每沒有指定P，S的值，那麼默認狀況下，NUMBER的P、S的值分別是多少呢？相信這個問題能問倒一大片DBA。 在以前，我遇到了一個問題，總結整理在「ORACLE NUMBER類型Scale爲0引起的問題」這篇博客當中，當時武斷的判斷「若是不指定p和s，NUMBER類型，它的默認精度值爲38, 默認的scale值爲0」，由於當時參考了官方文檔https://docs.oracle.com/cd/B28359_01/server.111/b28318/datatype.htm#CNCPT1832

 

 

固然文檔沒有錯誤，文檔應該是指在定義字段數據類型爲NUMBER時，指定了NUMBER類型的P值，可是沒有指定S的值，那麼Scale默認就是0，以下測試所示，當時應該是我本身沒有徹底理解文檔意思，固然文檔也有誤導的嫌疑。

 

 

SQL> drop table test;

Table dropped.

SQL> create table test(id number(38));

Table created.

SQL> insert into test

  2  select 123 from dual union all

  3  select 123.123 from dual;

2 rows created.

SQL> commit;

Commit complete.

SQL> select * from test;

        ID

----------

       123

       123

SQL> 

 

 

 

 

當在指定字段類型爲NUMBER時，若是P和S都不指定，那麼P和S又是什麼值呢？今天特地實驗驗證了一下，具體實驗過程以下：

 

 

SQL> drop table test;

Table dropped.

SQL> create table test(id  number, id1 number(38,4));

Table created.

SQL> insert into test                                      

  2  select 12, 12 from dual union all

  3  select 12.123456789, 12.123456789 from dual;

2 rows created.

SQL> commit;

Commit complete.

SQL> col id  for 999999999999.999999999999999999999999999999999999;

SQL> col id1 for 99999999999.9999999999999999999999999999999999999;

SQL> select * from test;

                                                ID                                                ID1

-------------------------------------------------- --------------------------------------------------

           12.000000000000000000000000000000000000           12.0000000000000000000000000000000000000

           12.123456789000000000000000000000000000           12.1235000000000000000000000000000000000

SQL> 

 

 

如上所示，當我插入上面兩條記錄後，發現若是不指定p和s，NUMBER類型，此時的Scale至少是9，咱們繼續測試，插入下面數據

 

 

SQL> insert into test

  2  select 12.123456789123456789123456789123456,

  3         12.123456789123456789123456789123456

  4  from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test;

                                                ID                                                ID1

-------------------------------------------------- --------------------------------------------------

           12.000000000000000000000000000000000000           12.0000000000000000000000000000000000000

           12.123456789000000000000000000000000000           12.1235000000000000000000000000000000000

           12.123456789123456789123456789123456000           12.1235000000000000000000000000000000000

 

 

以下所示，此時能夠看到Scale的值33了，那麼Scale的值是否能夠繼續變大呢？

 

 

 

 

SQL> insert into test

  2  select 12.123456789123456789123456789123456789123,

  3         12.123456789123456789123456789123456789123

  4  from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test;

                                                ID                                                ID1

-------------------------------------------------- --------------------------------------------------

           12.000000000000000000000000000000000000           12.0000000000000000000000000000000000000

           12.123456789000000000000000000000000000           12.1235000000000000000000000000000000000

           12.123456789123456789123456789123456000           12.1235000000000000000000000000000000000

           12.123456789123456789123456789123456789           12.1235000000000000000000000000000000000

 

 

以下截圖所示，插入的記錄爲12.123456789123456789123456789123456789123，可是顯示的值爲12.123456789123456789123456789123456789，總共爲38位，因爲格式化列的緣故，可能致使部分小數位沒有顯示，

 

 

 

咱們繼續測試，調整格式化列，咱們發現值變爲了12.12345678912345678912345678912345678912，總共40位了，Scale的值爲38了。這個是爲何呢？不是數字精度爲38，意味着最可能是38位嗎？

 

SQL> col id  for 999999999999.99999999999999999999999999999999999999999

SQL> col id1 for 99999999999.999999999999999999999999999999999999999999

SQL> select * from test;

                                                     ID                                                     ID1

------------------------------------------------------- -------------------------------------------------------

           12.00000000000000000000000000000000000000000           12.000000000000000000000000000000000000000000

           12.12345678900000000000000000000000000000000           12.123500000000000000000000000000000000000000

           12.12345678912345678912345678912345600000000           12.123500000000000000000000000000000000000000

           12.12345678912345678912345678912345678912000           12.123500000000000000000000000000000000000000

 

 

 

繼續其它測試，咱們發現Sacle的值會隨着小數點前面數字的變化而變化，以下所示：

 

SQL> insert into test

  2  select 123456789.123456789123456789123456789123456,

  3         123456789.123456789123456789123456789123456

  4  from dual;

1 row created.

SQL> commit;

Commit complete

SQL> insert into test

  2  select 123456789123.123456789123456789123456789123456,

  3         123456789123.123456789123456789123456789123456

  4  from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test;

                                                     ID                                                     ID1

------------------------------------------------------- -------------------------------------------------------

           12.00000000000000000000000000000000000000000           12.000000000000000000000000000000000000000000

           12.12345678900000000000000000000000000000000           12.123500000000000000000000000000000000000000

           12.12345678912345678912345678912345600000000           12.123500000000000000000000000000000000000000

           12.12345678912345678912345678912345678912000           12.123500000000000000000000000000000000000000

    123456789.12345678912345678912345678912300000000000    123456789.123500000000000000000000000000000000000000

 123456789123.12345678912345678912345678910000000000000 #######################################################

6 rows selected.

 

 

從上面測試能夠看出，Scale的值是變化的，跟數據值有關係，目前看來，小數點前的數字位數和小數點後的數字位數相加爲40（有時候又是39），爲了測試是否這個規律，我特地用下面案例測試一下

 

 

 

SQL> create table test2(id number);

Table created.

SQL> insert into test2

  2  select 0.123456789123456789123456789123456789123456789 from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> col id for 9999999999.9999999999999999999999999999999999999999999999;

SQL> select * from test2;

                                                        ID

----------------------------------------------------------

           .1234567891234567891234567891234567891235000000

SQL> insert into test2

  2  select 123456789.123456789123456789123456789123456789 from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test2;

                                                        ID

----------------------------------------------------------

           .1234567891234567891234567891234567891235000000

  123456789.1234567891234567891234567891230000000000000000

SQL> insert into test2

  2  select 123456789123.123456789123456789123456789123456789 from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test2;

                                                        ID

----------------------------------------------------------

           .1234567891234567891234567891234567891235000000

  123456789.1234567891234567891234567891230000000000000000

##########################################################

SQL>  col id for 9999999999999.9999999999999999999999999999999999999999999999;

SQL> select * from test2;

                                                           ID

-------------------------------------------------------------

              .1234567891234567891234567891234567891235000000

     123456789.1234567891234567891234567891230000000000000000

  123456789123.1234567891234567891234567891000000000000000000

SQL> insert into test2

  2  select 12345678912345.12345678912345678912345678912345 from dual;

1 row created.

SQL> commit;

Commit complete.

SQL> select * from test2;

                                                           ID

-------------------------------------------------------------

              .1234567891234567891234567891234567891235000000

     123456789.1234567891234567891234567891230000000000000000

  123456789123.1234567891234567891234567891000000000000000000

#############################################################

SQL> col id for 9999999999999999999.99999999999999999999999999999999999999999999;

SP2-0246: Illegal FORMAT string "9999999999999999999.99999999999999999999999999999999999999999999"

SQL> col id for 9999999999999999999.9999999999999999999999999999999999999999

SQL> select * from test2;

                                                           ID

-------------------------------------------------------------

                    .1234567891234567891234567891234567891235

           123456789.1234567891234567891234567891230000000000

        123456789123.1234567891234567891234567891000000000000

      12345678912345.1234567891234567891234567900000000000000

SQL> 

 

 

 

這個問題糾結了好久，不明白爲何是39或40，後面在Oracle Database SQL Reference 10g Release 2終於找到解釋了，以下所示：

 

p is the precision, or the total number of significant decimal digits, where the most

significant digit is the left-most nonzero digit, and the least significant digit is the

right-most known digit. Oracle guarantees the portability of numbers with

precision of up to 20 base-100 digits, which is equivalent to 39 or 40 decimal digits

depending on the position of the decimal point.

 

p是精度，或是有效十進制數的總位數。最大的有效數字是最左邊的非零數字，而最小有效位是最右邊的數字。 Oracle保證數字的可移植性

精度高達20 base-100 digits，至關於39位或40位十進制數字，取決於小數點的位置。