作者查詢 / celestialgod
作者 celestialgod 在 PTT [ Python ] 看板的留言(推文), 共75則
限定看板:Python
看板排序:
全部R_Language3057MATLAB995Statistics896PC_Shopping309Tech_Job186Soft_Job140Gossiping96Python75C_and_CPP45Boy-Girl24Aries23CareerPlan23BoardGame22DIABLO22GRE19DataScience18TOEIC18joke17hardware16NARUTO16ONE_PIECE16studyabroad15Hunter13Math13Tainan11Wanted11C_Chat10Lottery9PathofExile9SMSlife9StupidClown9BLEACH8China-Drama8CodeJob8LaTeX8Salary8TOEFL_iBT8Lifeismoney7LoL7Oversea_Job7Beauty6Economics6Golden-Award6HardwareSale6LCD6Hate5HomeTeach5JLPT5NBA5NSwitch5Isayama4movie4nb-shopping4WomenTalk4EAseries3home-sale3MacShop3PhD3trans_math3ArakawaCow2CFAiafeFSA2firsttime2Instant_Mess2Japandrama2NewAge2Notebook2sinica2Sub_CS2WarCraft2ask1Baseball1BigShiLin1CarShop1Chihlee1CrossStrait1friends1Hsinchu1java1kekkai1KUAS1Militarylife1MuscleBeach1Nangang1NCKU_ECO971nobunyaga1Office1Oni_soul1OverClocking1pal1Penpal1PingTung1PokeMon1Programming1PttLifeLaw1RFonline1specialman1SYSOP1TY_Research1VR1<< 收起看板(99)
2F推: .map05/21 16:19
9F推: var_dict = {“cat”: 0, “dog”: 1}05/22 19:25
10F→: df[‘var’] = df[‘var’].map(var_dict)05/22 19:25
11F→: var_dict = {“yes”: 0, “no”: 1}05/22 19:26
12F→: 這就可以把yes/no map to 1/005/22 19:26
15F推: df[‘var’] = df[‘var’].map({‘是’:1,’否’05/22 22:33
16F→: :0})05/22 22:33
4F→: 我之前是用Apache HTTP server 你只要建好目錄就可05/14 18:43
5F→: 以使用了05/14 18:43
6F→: https://tinyurl.com/5ubr2du705/14 18:45
8F→: 忘了說 Apache的話 只有下載功能而已 沒有publish05/16 18:29
17F推: 試試看Cython Linux只要裝好g++ pythondev windows05/17 13:05
18F→: 比較麻煩要裝VC++ 但是效能應該可以好很多05/17 13:05
19F→: https://tinyurl.com/yusbeekf05/17 13:07
20F→: 我那時候把Python改成Cython快了10倍以上05/17 13:07
1F推: AWS lambda, Azure Functions, Azure Logic App04/13 12:17
1F推: 不考慮numpy+Cython嗎?01/27 17:40
2F推: https://reurl.cc/k7lD3x01/27 17:41
3F→: 如果不考慮這麼複雜的方式 建議直接用armadillo重寫01/27 17:44
4F→: https://reurl.cc/dX76kg01/27 17:48
5F→: 雖然上面那篇是從matlab改C++但是numpy也同理01/27 17:49
16F推: 轉完之後可以01/28 10:17
4F→: https://reurl.cc/e9n3pM How about temp table?03/25 16:38
5F→: create a temp table in db and join it with03/25 16:39
6F→: your table03/25 16:39
7F→: 阿沒看到不插入資料下....03/25 16:40
4F→: 可以考慮armadillo, GSL等轉換會簡單很多09/28 13:52
1F→: sort by id, NO (descending) 然後group by id取08/27 01:14
2F→: 一筆08/27 01:14
3F→: https://pastebin.com/pcqXhLFt08/27 01:20
1F→: https://pastebin.com/GykDhbj208/25 22:42
5F→: 還可以直接試試看迴圈用numba.jit... 快很多XD08/26 00:56
1F→: https://goo.gl/1gYJgh03/25 19:42
2F→: https://goo.gl/Mv5nTX groupby字串的看起來還可以03/25 22:09
3F→: Test 403/25 22:10
4F→: by int或是numeric 就滿悲劇的XDD03/25 22:10
5F→: 所以你說Python一定比較快 恩... 應該還是不一定03/25 22:11
6F→: 而且data.table的測試指出pandas記憶體用太多03/25 22:12
7F→: 在dplyr, data.table沒爆的情況下,pandas爆了03/25 22:13
10F→: trace了一下issue,2E9列,pandas會爆掉那個已經fix03/25 22:18