To check whether there is any size mismatch between columns of two tables

Question

Hi,
Here i got two tables T and M where i am going to migrate data from T to M. But before migrating i need to check whether all the data in source table fits into destination table columns. The datatypes of all columns in source table T is of Varchar2 only as it is a temp table.
Ex :- Table 'T' (Source table) with columns
T_Lat Varchar2(50);
T_Amt Varchar2(50);
T_Cat Varchar2(50);
T_Vat Varchar2(50);
Now I have another table 'M'(Destination Table) with columns
M_Lat Varchar2(50);
M_Amt varchar2(25);
M_Cat date;
M_Vat number;
Now my task is I have to do a prevalidation of the data in 'T' that whether all the data in those columns will suit for the destination table columns respectively (to check whether there is any size mismatch). This check should be done dynamically.
For suppose, in T_Amt(source column of T table) if text is abt 50 characters, it cant fit M_Amt(destination column of M table). In this case it should throw an error indicating that destination column size is less for the source column.
Note:- There is no unique mapping column for these two tables and there are about 400 columns in the source table to be validate
I think it can be done using arrays or plsql tables.
Can any one help in this regard.

Answer 1

>
Now my task is I have to do a prevalidation of the data in 'T' that whether all the data in those columns will suit for the destination table columns respectively (to check whether there is any size mismatch). This check should be done dynamically.
>
Just because the source table T_Amt column is defined as 50 doesn't mean any of the data is really that long. So the data itself needs to be checked. That is just what a simple query can do.
See my answer in this thread Posted: Jul 25, 2012 1:14 PM
Re: Help in Execute Immediate - Invalid relational Operator
Here is the modified sample query and results for a query of a clone of the EMP table with some modified data
select count(*) cnt,
       -- job column
       sum(case when job is null then 1 else 0 end) job_nul,
       sum(case when job = 'SALESMAN' then 1 else 0 end) job_salesman,
       min(length(job)) job_minlength, max(length(job)) job_maxlength,
       min(job) job_min, max(job) job_max,
       -- hiredate colulmn
       sum(case when hiredate is null then 1 else 0 end) hiredate_nul,
       min(hiredate) hiredate_min,
       max(hiredate) hiredate_max
        from emp1
CNT JOB_NUL JOB_SALESMAN JOB_MINLENGTH JOB_MAXLENGTH JOB_MIN JOB_MAX  HIREDATE_NUL HIREDATE_MIN HIREDATE_MAX
14    2       4            5             9             ANALYST SALESMAN 0            9/28/0001    12/3/9999With one query and ONE pass thru the table I was able to get the COUNTs, the MIN and MAX values for each column and the MIN and MAX lengths of the VARCHAR2 columns.
There are 14 total records, 2 where JOB is null, 4 where the JOB is SALESMAN. The MIN length of the JOB data is 5 and the MAX is 9.
Look at the date values. The results tell me I have some problem data.
The length data tells me if I try to put the JOB data into another table I need to define the length as at least 9 or it won't fit.
For your use case you might find that all of the data in the T_Lat column is shorter than 26 and will actually fit into the M_Lat target table column.