对于大型数据处理的“庖丁解牛” | /cn/2007/08/deal-with-large-data-using-r/

yihui 2022-12-16 05:38:12

对于大型数据处理的“庖丁解牛”

https://yihui.org/cn/2007/08/deal-with-large-data-using-r/

8 Comments

giscus-bot 2022-12-16 05:38:35

访客 *五脚星* @ 2007-08-11 12:41:19 写道：

😎 sofa!

giscus-bot 2022-12-16 05:38:36

Guest *rtist* @ 2007-08-13 09:57:29 originally posted:

Did you try something on the netflix data?
I got many troubles on it... for such a big data set.
Creating a table in MySQL seems to need more than a day's harddisk noisy "clicking"...

yihui 2022-12-16 05:39:24

I haven't tried that data yet... But I don't think my method above will be effective for netfix data, because I find there're only some three or four variables. However, ignoring the customer id (they are randomly-assigned) might save some time.

BTW, I didn't check the data carefully... It seems that there are only two predictors? ❓ MovieID & Date?

Originally posted on 2007-08-13 15:22:13

giscus-bot 2022-12-16 05:38:37

Guest *rtist* @ 2007-08-13 19:48:12 originally posted:

My question is naive: how will you make a databases 'efficiently' from those 17770 text files, before using any of your stuffs? There are 3~4 predictors: movieID, customerID, ratingDate, and movieReleaseDate. But the reference method only used movieID and customerID. The customerID is an important predictor.

yihui 2022-12-16 05:39:25

It took me less than an hour to merge these 17770 files into a single csv file (using my Windows XP with 512M memory). You just read these files and write them into a MySQL database? I didn't try it.

Originally posted on 2007-08-14 09:24:01

giscus-bot 2022-12-16 05:38:37

Guest *rtist* @ 2007-08-14 20:02:58 originally posted:

I have problems in replying under Ubuntu. So I wrote it down at http://piggy-rtist.spaces.live.com/blog/cns!F0F3CFF4E5F183CC!345.entry

giscus-bot 2022-12-16 05:38:38

访客 *littlesas* @ 2008-11-23 20:25:51 写道：

谢大哥，你好厉害，我在学R，向你看齐。

yihui 2022-12-16 05:39:26

一般一般……别向我看齐，我又不是什么标杆人物……

——原帖发布于 2008-11-24 14:45:13

giscus-bot 2022-12-16 05:38:39

访客 *myli* @ 2009-01-22 12:27:22 写道：

无意中看到你的博客，收获颇多。
我不是统计出身，但因在数据公司打混，所以很想把统计学精了，以便用于工作中。
真希望能和你多交流哈， 👍

@licw58

giscus-bot 2022-12-16 05:38:39

访客 *king64* @ 2009-03-24 06:38:32 写道：

Yihui，看你的东西长学问了！

如何是Access格式的数据库如何做？我尝试修改了下面语句，但不成功！
channel = odbcConnect("Text Files")
print(sqlColumns(channel, "x.csv")[, c(1, 3, 4, 6)])

yihui 2022-12-16 05:39:26

RODBC默认有对Access数据库的支持，参见ocbcConnectAccess函数；本文的文本数据是特殊“数据库”，因此我自己建立了特殊的ODBC连接。

——原帖发布于 2009-03-24 11:08:39

giscus-bot 2022-12-16 05:38:40

访客 *Stivensen295* @ 2012-08-20 08:59:09 写道：

我从大哥这里学习了很多R的应用，受益匪浅。伴随大哥的资料也顺利完成了研究生学业。愿大哥有更多精品问世。

yihui 2022-12-16 05:39:27

谢谢！大家都在陆续毕业，就我至今还在学校混日子啊。

——原帖发布于 2012-08-21 03:31:26

giscus-bot 2022-12-16 05:39:28

访客 *zilhua 裴* @ 2012-08-29 09:23:08 写道：

新人，学习中