2013-10-07 137 views
2

搜索了兩天至目前爲止無效的解決方案。R:如何彙總數據但保留來自非彙總列的信息?

我從不同的觀察點觀察鳥類。觀察員記下物種,他們在哪裏見過它們,以及記錄多久。

現在發生的情況是,從不同的角度來看,觀測是從同一個地區得到的,但我們只想處理一個地區的每個物種的最大值。

因此,首先,我通過觀察點,品種和區域彙總的數據,並總結出了時間。

dt.agg <- aggregate(time ~ observp + species + time, dt, sum) 

UUPS:完全地錯誤的命令:

應該是:

dt.agg <- aggregate(time ~ observp + species + area, dt, sum) 



    observp species area time 
1  1a Rm A1  43.878488 
2  1c Rm A1  296.152707 
3  2 Rm A1  29.546790 
4  1a Swm A1  34.127713 
5  1b Swm A1  11.076880 
6  2 Swm A1   8.771703 

這個工作正常。但現在,我只需要一個地區物種的最大時間值,但是我還需要知道從哪個觀測點獲得這些數字。

在我的示例中,第2行應保留在A1中的Rm,而第1行和第3行應該被刪除。這同樣適用於第4行(保持)和5 + 6(下降)

當我剛做的品種和麪積隨着時間的推移和最大其它集合,爲觀察點的信息丟失。

有人可以告訴我一種方法來實現嗎?

乾杯

貝恩德

(現在新的帳戶,並沒有信譽..謝謝...谷歌!)

附:請隨時給這個問題一個更好的標題

UPDATE: 試圖張貼dput(head(dt,100)) - 樣本建議。原始數據集有超過1300行。希望這就是你想要的。

structure(list(species = structure(c(3L, 3L, 3L, 5L, 5L, 5L, 
5L, 5L, 5L, 5L, 5L, 5L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 5L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 5L, 5L, 
5L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 5L, 
5L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 5L, 5L, 5L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 5L, 5L, 5L, 5L, 5L, 3L, 3L, 3L, 5L, 5L, 5L, 
3L, 3L, 3L, 3L, 3L, 3L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L), .Label = c("Bf", 
"Gr", "Rm", "Row", "Swm", "Wf", "Wsb", "Wst", "Ww"), class = "factor"), 
    area = structure(c(35L, 19L, 34L, 34L, 32L, 19L, 34L, 35L, 
    10L, 36L, 10L, 14L, 13L, 25L, 27L, 28L, 34L, 19L, 14L, 14L, 
    34L, 1L, 12L, 13L, 15L, 3L, 3L, 34L, 34L, 34L, 14L, 14L, 
    13L, 13L, 1L, 1L, 1L, 11L, 1L, 8L, 21L, 22L, 22L, 9L, 9L, 
    9L, 5L, 9L, 3L, 22L, 27L, 26L, 21L, 26L, 21L, 27L, 3L, 9L, 
    20L, 20L, 9L, 26L, 34L, 30L, 3L, 2L, 3L, 4L, 20L, 3L, 37L, 
    16L, 17L, 18L, 14L, 35L, 34L, 34L, 34L, 36L, 4L, 4L, 3L, 
    3L, 17L, 17L, 38L, 36L, 10L, 38L, 36L, 10L, 38L, 37L, 35L, 
    30L, 16L, 15L, 17L, 5L), .Label = c("A1", "A10", "A11", "A12", 
    "A13", "A14", "A15", "A16", "A17", "A18", "A2", "A3", "A4", 
    "A5", "A6", "A7", "A8", "A9", "O1", "O10", "O11", "O12", 
    "O13", "O14", "O15", "O16", "O17", "O18", "O19", "O2", "O20", 
    "O21", "O22", "O3", "O4", "O5", "O7", "O8", "O9"), class = "factor"), 
    observp = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
    1L), .Label = c("1a", "1b", "1c", "2", "3", "4"), class = "factor"), 
    time = c(36.37086972, 2.730715967, 1.891286914, 3.782573827, 
    4.496276059, 5.461431934, 18.91286914, 13.22577081, 5.823001976, 
    5.392743201, 3.882001317, 16.97305991, 6.094384821, 5.274262222, 
    5.462035947, 2.089427691, 7.565147654, 21.84572774, 25.45958986, 
    16.97305991, 7.565147654, 4.875387532, 8.885792099, 4.062923214, 
    6.636122805, 7.038317277, 10.55747592, 7.565147654, 7.565147654, 
    3.782573827, 25.45958986, 25.45958986, 12.18876964, 12.18876964, 
    19.50155013, 19.50155013, 9.750775065, 39.20627398, 4.875387532, 
    6.423076843, 2.436283538, 1.823249104, 1.823249104, 16.72889022, 
    41.82222555, 33.45778044, 12.30932064, 117.1022315, 3.519158639, 
    1.823249104, 27.31017974, 11.11346598, 4.872567077, 11.11346598, 
    4.872567077, 5.462035947, 3.519158639, 16.72889022, 14.86012871, 
    8.916077225, 25.09333533, 22.22693195, 3.782573827, 5.184879322, 
    10.55747592, 8.509038411, 10.55747592, 17.70988435, 5.944051483, 
    3.519158639, 17.69229328, 34.70586347, 5.966017168, 3.092236431, 
    2.828843318, 6.612885403, 3.782573827, 3.782573827, 7.565147654, 
    5.392743201, 17.70988435, 17.70988435, 3.519158639, 2.346105759, 
    11.93203434, 11.93203434, 2.386548395, 0.898790534, 0.64700022, 
    2.386548395, 0.898790534, 0.64700022, 2.684866944, 6.634609979, 
    1.239916013, 1.944329746, 3.2536747, 3.732819078, 6.711769315, 
    2.307997621)), .Names = c("species", "area", "observp", "time" 
), row.names = c(NA, 100L), class = "data.frame") 
+0

嗨,你可以提供'dput(dt)'的結果,或者它的一個子集,如果它太大,就像'dput(head(dt,100))'?另外,您嘗試過的第二個「聚合」是什麼? – Frank

+0

Bernd,請使用'dput(yourobject)'或'dput(head(yourobject))'將您的數據(或假數據)樣本添加到您的問題中。 [這篇文章](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example)是如何做到這一點的規範資源。 – SlowLearner

+0

我不是一段時間(帳戶混淆)。標記的解決方案爲我工作。沒有時間檢查其他人。感謝您的親切幫助! –

回答

2

您可能還會看看另一個base函數,by。輸出是一個列表,其中每個元素是INDICES的不同組合的結果。

bb <- by(data = df, INDICES = list(df$species, df$area), function(x) x[which.max(x$time), ]) 
bb 
# : Rm 
# : A1 
# observp species area  time 
# 2  1c  Rm A1 296.1527 
# -------------------------------------------------------------------- 
# : Swm 
# : A1 
# observp species area  time 
# 4  1a  Swm A1 34.12771 

如果要將列表轉換爲data.frame

df2 <- do.call(rbind, bb) 
df2 
# observp species area  time 
# 2  1c  Rm A1 296.15271 
# 4  1a  Swm A1 34.12771 

另一種選擇:

library(plyr) 
ddply(.data = df, .variables = .(species, area), subset, 
    time == max(time)) 
+0

這一個似乎是我第一次聚合的結果。我得到了同樣的285行,就像我簡單的第二次聚集的物種和麪積一樣,但是這次有適當的觀察點保存到數據中。將明天驗證,現在太晚了......迄今爲止,還有很多感謝! –

0

一個例子是

stulevel_agg_2 <- stulevel[, list(a1=mean(ability, na.rm = TRUE), a2=last(school, na.rm=T)),by = grade]

a1, a2是新列名稱。 last可以取組中的最後一個元素,但首先需要加載xts