jar file too large...

Started by Septisin, 25. May 2008, 10:09:53

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Septisin

I`m trying to create a large French/Czech dictionary (375 000 entries, 23 000kb in txt file). Until JarCreator everything works fine. DictionaryGeneration had created about 5700 csv files, 130 mb together, but after then JarCreator tried to create a too large jar file, it always stopped with a message that there is not enough disk space. At that moment was the size of the jar file about 70 Gb... My cell phone is Nokia 6110 Navigator with no restriction of jar file size. Whats wrong? Thanks for your help!

Gert

When DictionaryGeneration creates 130 MB out of a 23 MB file, then probably something went wrong with the creation of the indexes. Probably huge indexes had been created. Could you check this ? If yes, there are ways to reduce the size of the index.

I did not completely understand the problem of JarCreator. A little bit simplified JarCreator just takes the files that are created from DictionaryGeneration and adds them to the 'empty' JAR file. The resulting size will be less than the size of the files generated from DictionaryGeneration because the JAR files are compressed.

70 GB of file size ? Or 70 MB ? Could you check that you have separate directories (folders) for input and output ?

When the files are correctly built, then I think the result should run fine on a Nokia 6110.

Gert


Septisin

DictionaryGeneration created 3800 index csv files/97 Mb. It`s really 70 Gb and it never finished because it filled up my C: disk. I`ll check my folder structure later - today is unfortunately  :P a family day... Thanks for your quick reply! I`ll be back as soon as possible!

Gert

No need to hurry ;)

You certainly need to check what you want to have indexed - 97 MB index file could come from indexing full sentences like 'expressions' (which is certainly not what you want; maybe indexing the words of the sentences could be an option).

Gert

Septisin

You are right, when i was copying the dictionary (Quickmacro/Text Capture - by the way the only program i found which supports the eastern european fonts and makes usable text formating) i forgot to switch off expressions (there is lot of them and they are useless in such type of dictionary) and i didn't repair it, becouse i thought, that the only inconvenience will be a bigger final file (i didn't expect that it will be gigabytes, if it's the real reason of seething of the jar file - now i'm at work i'll check the folder structure this evening). So i'll repair the txt file and I'll give news soon. Thanks for your help!

Gert