将多个 dataframe 列 int64 dtype 转换为等效于 oracle 的 NUMBER(20,3) 数据类型

Question

I am trying to insert dataframe in oracle DB table but some of columns in table has datatype as NUMBER(20,3)

I am assuming this is float.

How can convert my column having datatype as int64 to NUMBER(20,3) equivalent as it is giving invalid identifier error .

Answer 1

number(20, 3) in Oracle means that column accepts numeric value which have 20 digits in total, and 3 out of those 20 are reserved for decimals. For example:

SQL> create table test (id number, value number(20, 3));

Table created.

Let's insert some values:

SQL> insert into test (id, value) values (1, 12.43);

1 row created.

SQL> insert into test (id, value) values (2, 12.4356);

1 row created.

SQL> insert into test (id, value) values (3, 12345678901234567.987);

1 row created.

Table contents: no surprises:

SQL> select * from test;

        ID      VALUE
---------- ----------
         1      12,43
         2     12,436
         3 1,2346E+16

What if you insert more than 3 decimal digits? Oracle will perform round :

SQL> insert into test (id, value) values (4, 12.4327);

1 row created.

---

But, you reported invalid identifier error. That has nothing to do with contents you're inserting, but invalid column name. Have a look: I "renamed" the id column to idxyz and Oracle raised error you mentioned:

SQL> insert into test (idxyz, value) values (5, 123);
insert into test (idxyz, value) values (5, 123)
                  *
ERROR at line 1:
ORA-00904: "IDXYZ": invalid identifier


SQL>

---

Therefore, check statement you use, looks like you misspelled one (or more) column names. Mind double quotes around column names (in Oracle, we usually don't use them) and - if you do have double quotes - mind letter case as well.

---

Oh, yes: how to convert int64 to number(20, 3) ? No idea, I don't use Python.