从Oracle中的表将大量数据导出到平面文件

Question

I am using Oracle 11g. My requirement is to compare the data of two different db's. There are around 350 tables in each db and out of these 350 tables, approx 40 tables have more than 1 million records. For data comparison, I wrote one perl script to compare using hash and tested with few files. Also, tried with unix awk command to check the performance and asked this forum on unix solution and got excellent help.

Now my problem is to find out the best way to extract data from Tables to files.
Both db's have same number of tables and each table will have same number of columns in both db ie the layout in both the db's is exactly the same.
options which i think and searched are
1) using sqlloader - I think performance will be bad in this case
2) using data pump - Not sure if i can extract few set of columns via sql by using data pump and load into text file
3) using bulk collect -- same as above. Is it possible to extract each table and from each table set of columns. if yes, how can it be done. Also what would be the performance.
4) sqlplus or anything else. I cannot download any software for this on my machine.

Basic sql for selecting set of columns from each table for both the db's can be done easily. I am looking at the best approach to export data into flat files.
Please suggest

Answer 1

You can do it in the database using DBMS_COMPARISON.

http://docs.oracle.com/cd/E11882_01/appdev.112/e25788/d_comparison.htm#CHDHEFFJ

Answer 2

The fundamental approach that is the least expensive (to the developer for sure) is to compare sets of things rather than string evaluations of files. Nothing can compare sets of things faster (and in less code) than the database itself. The use of database links and a wise use of the MINUS and INTERSECT operators is very powerful means towards this end.

Answer 3

Try using the below SQL this should be the fastest approach as you will be working inside the database. Access the table over the DB link in other database.

select * from ( ( select * from Table_In_Schema1 minus select * from Table_In_Schema2) union all ( select * from Table_In_Schema2 minus select * from Table_In_Schema1) )