Question

I am using Derby In-Memory DB. I need to perform some data loading from csv files in the beginning. For now, it takes about 25 seconds to load all the csv files into their tables. I hope the time can be reduced. Due to the data files are not very large actually.

我所做的是利用从失败中起的内在程序。

{CALL SYSCS_UTIL.SYSCS_IMPORT_TABLE (?,?,?, , , " , UTF-8 ,1 )} or
{CALL SYSCS_UTIL.SYSCS_IMPORT_TABLE (?,?,?, , , " , UTF-8 ,0 )}

唯一的特殊事项是,一个表格中的数据有时被分成许多小的卷宗。因此,我不得不逐一装载。如果我能够把他们合并在一起,我就进行了测试,只有16秒。然而,由于用户需要,我无法消除这一特征。

Is there anything I can do to reduce the time of loading data? Should I disable log or write some user-defined function/procedure or any other tune can be done? Any advice will be fine.

感谢!

Answer 1

Use H2 instead of Derby, and use the CSVREAD feature. If that s still too slow, see the fast import optimization, or use the CSV tool directly (without using a database). Disclaimer: I wrote the CSV support for H2.

友情链接