【独家发布】Data Manipulation with R

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

https://www.packtpub.com/big-data-and-business-intelligence/data-manipulation-r

Data Manipulation with R

January 2014

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏4 回帖

关键词：manipulation ulation ATION Data With 2014

1. ##################
   
2. # Code Snipped-1
   
3. ##################
   
4. 
   
5. # Constant
   
6. 2
   
7. "July"
   
8. NULL
   
9. NA
   
10. NaN
    
11. Inf
    
12. 
    
13. # Object can be created from existing object
    
14. # to make the result reproducible mean every time we run the following code we will get same results # we need to set a seed value
    
15. set.seed(123)
    
16. rnorm(9)+runif(9)
    
17. 
    
18. ##################
    
19. # Code Snipped-2
    
20. ##################
    
21. 
    
22. # Storing R object into a variable and then see the mode
    
23. 
    
24. num.obj <- seq(from=1,to=10,by=2)
    
25. mode(num.obj)
    
26. logical.obj<-c(TRUE,TRUE,FALSE,TRUE,FALSE)
    
27. mode(logical.obj)
    
28. character.obj <- c("a","b","c")
    
29. mode(character.obj)
    
30. 
    
31. 
    
32. ##################
    
33. # Code Snipped-3
    
34. ##################
    
35. 
    
36. # R object containing both numeric and logical element
    
37. xz <- c(1, 3, TRUE, 5, FALSE, 9)
    
38. xz
    
39. mode(xz)
    
40. # R object containing character, numeric and logical elements
    
41. xw <- c(1,2,TRUE,FALSE,"a")
    
42. xw
    
43. mode(xw)
    
44. 
    
45. ##################
    
46. # Code Snipped-4
    
47. ##################
    
48. 
    
49. num.obj <- seq(from=1,to=10,by=2)
    
50. logical.obj<-c(TRUE,TRUE,FALSE,TRUE,FALSE)
    
51. character.obj <- c("a","b","c")
    
52. is.numeric(num.obj)
    
53. is.logical(num.obj)
    
54. is.character(num.obj)
    
55. 
    
56. ##################
    
57. # Code Snipped-5
    
58. ##################
    
59. 
    
60. mode(mean)
    
61. # Also we can test whether "mean" is function or not as follows
    
62. is.function(mean)
    
63. 
    
64. ##################
    
65. # Code Snipped-6
    
66. ##################
    
67. 
    
68. num.obj <- seq(from=1,to=10,by=2)
    
69. logical.obj<-c(TRUE,TRUE,FALSE,TRUE,FALSE)
    
70. character.obj <- c("a","b","c")
    
71. class(num.obj)
    
72. class(logical.obj)
    
73. class(character.obj)
    
74. 
    
75. ##################
    
76. # Code Snipped-7
    
77. ##################
    
78. 
    
79. # Output omitted due to space limitation
    
80. num.obj <- seq(from=1,to=10,by=2)
    
81. set.seed(1234) # To make the matrix reproducible
    
82. mat.obj <- matrix(runif(9),ncol=3,nrow=3)
    
83. mode(num.obj)
    
84. mode(mat.obj)
    
85. class(num.obj)
    
86. class(mat.obj)
    
87. # prints a numeric object
    
88. print(num.obj)
    
89. #prints a matrix object
    
90. print(mat.obj)
    
91. 
    
92. ##################
    
93. # Code Snipped-8
    
94. ##################
    
95. 
    
96. character.obj <- c("a","b","c")
    
97. character.obj
    
98. is.factor(character.obj)
    
99. 
    
100. # Converting character object into factor object using as.factor()
     
101. factor.obj <- as.factor(character.obj)
     
102. factor.obj
     
103. is.factor(factor.obj)
     
104. mode(factor.obj)
     
105. class(factor.obj)
     
106. 
     
107. ##################
     
108. # Code Snipped-9
     
109. ##################
     
110. 
     
111. # creating vector of numeric element with "c" function
     
112. num.vec <- c(1,3,5,7)
     
113. num.vec
     
114. mode(num.vec)
     
115. class(num.vec)
     
116. is.vector(num.vec)
     
117. 
     
118. ##################
     
119. # Code Snipped-10
     
120. ##################
     
121. 
     
122. # Vector with mixed elements
     
123. num.char.vec <- c(1,3,"five",7)
     
124. num.char.vec
     
125. mode(num.char.vec)
     
126. class(num.char.vec)
     
127. is.vector(num.char.vec)
     
128. 
     
129. ##################
     
130. # Code Snipped-11
     
131. ##################
     
132. 
     
133. # combining multiple vectors
     
134. comb.vec <- c(num.vec,num.char.vec)
     
135. mode(comb.vec)
     
136. 
     
137. # creating named vector
     
138. named.num.vec <- c(x1=1,x2=3,x3=5)
     
139. named.num.vec
     
140. 
     
141. ##################
     
142. # Code Snipped-12
     
143. ##################
     
144. 
     
145. # vector of single element
     
146. unit.vec <- 9
     
147. is.vector(unit.vec)
     
148. 
     
149. 
     
150. ##################
     
151. # Code Snipped-13
     
152. ##################
     
153. 
     
154. # creating a vector of numbers
     
155. # and then convert it to logical and character
     
156. numbers.vec <- c(-3,-2,-1,0,1,2,3)
     
157. numbers.vec
     
158. num2char <- as.character(numbers.vec)
     
159. num2char
     
160. num2logical <- as.logical(numbers.vec)
     
161. num2logical
     
162. 
     
163. # creating character vector
     
164. #and then convert it to numeric and logical
     
165. char.vec <- c("1","3","five","7")
     
166. char.vec
     
167. char2num <- as.numeric(char.vec)
     
168. char2num
     
169. char2logical <- as.logical(char.vec)
     
170. char2logical
     
171. 
     
172. # logical to character conversion
     
173. logical.vec <- c(TRUE, FALSE, FALSE,  TRUE,  TRUE)
     
174. logical.vec
     
175. logical2char <- as.character(logical.vec)
     
176. logical2char
     
177. 
     
178. ##################
     
179. # Code Snipped-14
     
180. ##################
     
181. 
     
182. # creating a vector and accessing elements
     
183. vector1 <- c(1,3,5,7,9)
     
184. vector1
     
185. # accessing second elements of "vector1"
     
186. vector1[2]
     
187. 
     
188. # accessing three elements starting from second element
     
189. vector1[2:4]
     
190. 
     
191. # another way of creating vector. Here "from" is the starting point of the vector and "to" is the
     
192. # end point of the vector and "by" is increment
     
193. vector2 <- seq(from=2, to=10, by=2)
     
194. is.vector(vector2)
     
195. 
     
196. ##################
     
197. # Code Snipped-15
     
198. ##################
     
199. 
     
200. #creating factor variable with only one argument with factor()
     
201. factor1 <- factor(c(1,2,3,4,5,6,7,8,9))
     
202. factor1
     
203. levels(factor1)
     
204. labels(factor)
     
205. labels(factor1)
     
206. 
     
207. #creating factor with user given levels to display
     
208. factor2 <- factor(c(1,2,3,4,5,6,7,8,9),labels=letters[1:9])
     
209. factor2
     
210. levels(factor2)
     
211. labels(factor2)
     
212. 
     
213. ##################
     
214. # Code Snipped-16
     
215. ##################
     
216. 
     
217. # creating numeric factor and trying to find out mean
     
218. num.factor <- factor(c(5,7,9,5,6,7,3,5,3,9,7))
     
219. num.factor
     
220. mean(num.factor)
     
221. 
     
222. ##################
     
223. # Code Snipped-17
     
224. ##################
     
225. 
     
226. num.factor <- factor(c(5,7,9,5,6,7,3,5,3,9,7))
     
227. num.factor
     
228. 
     
229. #as.numeric() function only returns internal values of the factor
     
230. as.numeric(num.factor)
     
231. # now see the levels of the factor
     
232. levels(num.factor)
     
233. as.character(num.factor)
     
234. 
     
235. # now to convert the "num.factor" to numeric there are two method
     
236. # method-1:
     
237. mean(as.numeric(as.character(num.factor)))
     
238. 
     
239. # method-2:
     
240. mean(as.numeric(levels(num.factor)[num.factor]))
     
241. 
     
242. ##################
     
243. # Code Snipped-18
     
244. ##################
     
245. 
     
246. #creating vector of different variables and then create data frame
     
247. var1 <- c(101,102,103,104,105)
     
248. var2 <- c(25,22,29,34,33)
     
249. var3 <- c("Non-Diabetic", "Diabetic", "Non-Diabetic", "Non-Diabetic", "Diabetic")
     
250. var4 <- factor(c("male","male","female","female","male"))
     
251. 
     
252. # now we will create data frame using two numeric vector one
     
253. # character vector and one factor
     
254. diab.dat <- data.frame(var1,var2,var3,var4)
     
255. diab.dat
     
256. 
     
257. ##################
     
258. # Code Snipped-19
     
259. ##################
     
260. 
     
261. #class of each column before creating data frame
     
262. class(var1)
     
263. class(var2)
     
264. class(var3)
     
265. class(var4)
     
266. 
     
267. ##################
     
268. # Code Snipped-20
     
269. ##################
     
270. 
     
271. # class of each column after creating data frame
     
272. class(diab.dat$var1)
     
273. class(diab.dat$var2)
     
274. class(diab.dat$var3)
     
275. class(diab.dat$var4)
     
276. 
     
277. # now create the data frame specifying as.is=TRUE
     
278. diab.dat.2 <- data.frame(var1,var2,var3,var4,stringsAsFactors=FALSE)
     
279. diab.dat.2
     
280. class(diab.dat.2$var3)
     
281. 
     
282. ##################
     
283. # Code Snipped-21
     
284. ##################
     
285. 
     
286. # data frame to matrix conversion
     
287. mat.diab <- as.matrix(diab.dat)
     
288. mat.diab
     
289. class(mat.diab)
     
290. mode(mat.diab)
     
291. 
     
292. # matrix multiplication is not possible
     
293. # with this newly created matrix
     
294. 
     
295. t(mat.diab) %*% mat.diab
     
296. 
     
297. # creating a matrix with numeric elements only
     
298. # To produce the same matrix over time we set a seed value
     
299. set.seed(12345)
     
300. num.mat <- matrix(rnorm(9),nrow=3,ncol=3)
     
301. num.mat
     
302. class(num.mat)
     
303. mode(num.mat)
     
304. 
     
305. # matrix multiplication
     
306. t(num.mat) %*% num.mat
     
307. 
     
308. ##################
     
309. # Code Snipped-22
     
310. ##################
     
311. 
     
312. mat.array=array(dim=c(2,2,3))
     
313. 
     
314. # To produce the same results over time we set a seed value
     
315. set.seed(12345)
     
316. mat.array[,,1]<-rnorm(4)
     
317. mat.array[,,2]<-rnorm(4)
     
318. mat.array[,,3]<-rnorm(4)
     
319. mat.array
     
320. 
     
321. ##################
     
322. # Code Snipped-23
     
323. ##################
     
324. 
     
325. var1 <- c(101,102,103,104,105)
     
326. var2 <- c(25,22,29,34,33)
     
327. var3 <- c("Non-Diabetic", "Diabetic", "Non-Diabetic", "Non-Diabetic", "Diabetic")
     
328. var4 <- factor(c("male","male","female","female","male"))
     
329. diab.dat <- data.frame(var1,var2,var3,var4)
     
330. 
     
331. mat.array=array(dim=c(2,2,3))
     
332. set.seed(12345)
     
333. 
     
334. mat.array[,,1]<-rnorm(4)
     
335. mat.array[,,2]<-rnorm(4)
     
336. mat.array[,,3]<-rnorm(4)
     
337. 
     
338. # creating list
     
339. obj.list <- list(elem1=var1,elem2=var2,elem3=var3,elem4=var4,elem5=diab.dat,elem6=mat.array)
     
340. obj.list
     
341. 
     
342. 
     
343. ##################
     
344. # Code Snipped-24
     
345. ##################
     
346. 
     
347. missing_dat <- data.frame(v1=c(1,NA,0,1),v2=c("M","F",NA,"M"))
     
348. missing_dat
     
349. is.na(missing_dat$v1)
     
350. is.na(missing_dat$v2)
     
351. any(is.na(missing_dat))

复制代码

1. ##################
   
2. # Code Snipped-1
   
3. ##################
   
4. 
   
5. # Before running the following command we need to set the data
   
6. # location using setwd(). For example setwd("d:/chap2").
   
7. 
   
8. anscombe <- read.csv("CSVanscombe.csv",skip=2)
   
9. 
   
10. ##################
    
11. # Code Snipped-2
    
12. ##################
    
13. 
    
14. # import csv file that contains both numeric and character variable
    
15. # firstly using default and then using stringsAsFActors=FALSE
    
16. 
    
17. iris_a <- read.csv("iris.csv")
    
18. str(iris_a)
    
19. 
    
20. ##################
    
21. # Code Snipped-3
    
22. ##################
    
23. 
    
24. # Now using stringsAsFactors=FALSE
    
25. iris_b <- read.csv("iris.csv",stringsAsFactors=F)
    
26. str(iris_b)
    
27. 
    
28. ##################
    
29. # Code Snipped-4
    
30. ##################
    
31. 
    
32. iris_semicolon <- read.csv("iris_semicolon.csv",stringsAsFactors=FALSE,sep=";")
    
33. str(iris_semicolon)
    
34. 
    
35. ##################
    
36. # Code Snipped-5
    
37. ##################
    
38. 
    
39. anscombe_tab <- read.csv("anscombe.txt",sep="\t")
    
40. anscombe_tab_2 <- read.table("anscombe.txt",header=TRUE)
    
41. 
    
42. ##################
    
43. # Code Snipped-6
    
44. ##################
    
45. # Calling xlsx library
    
46. library(xlsx)
    
47. # importing xlsxanscombe.xlsx
    
48. anscombe_xlsx <- read.xlsx2("xlsxanscombe.xlsx",sheetIndex=1)
    
49. 
    
50. ##################
    
51. # Code Snipped-7
    
52. ##################
    
53. 
    
54. # loading robjects.RData file
    
55. load("robjects.RData")
    
56. 
    
57. # to see whether the objects are imported correctly
    
58. objects()
    
59. 
    
60. ##################
    
61. # Code Snipped-8
    
62. ##################
    
63. 
    
64. library(foreign)
    
65. iris_stata <- read.dta("iris_stata.dta")
    
66. 
    
67. ##################
    
68. # Code Snipped-9
    
69. ##################
    
70. 
    
71. # creating an R objects whose value is "datamanipulation"
    
72. char.obj <- "datamanipulation"
    
73. 
    
74. # creating a factor variable by extracting each single letter from the
    
75. # character string. To extract each single letter substring() function
    
76. # has been used. Note: nchar() function give number of character count
    
77. # in an character type R object
    
78. factor.obj <- factor(substring(char.obj,1:nchar(char.obj),1:nchar(char.obj)),levels=letters)
    
79. 
    
80. # Displaying levels of the factor variable
    
81. levels(factor.obj)
    
82. 
    
83. # Displaying the data using table() function
    
84. table(factor.obj)
    
85. 
    
86. 
    
87. ##################
    
88. # Code Snipped-10
    
89. ##################
    
90. 
    
91. # re-creating factor variable from existing factor variable
    
92. factor.obj1 <- factor(factor.obj)
    
93. 
    
94. # Displaying levels of the new factor variable
    
95. levels(factor.obj1)
    
96. 
    
97. # displaying data using table() function
    
98. table(factor.obj1)
    
99. factor.obj1
    
100. 
     
101. 
     
102. ##################
     
103. # Code Snipped-11
     
104. ##################
     
105. 
     
106. # creating a numeric variable by taking 100 random numbers
     
107. # from normal distribution
     
108. set.seed(1234) # setting seed to reproduce the example
     
109. numvar <- rnorm(100)
     
110. 
     
111. # creating factor variable with 5 distinct category
     
112. 
     
113. num2factor <- cut(numvar,breaks=5)
     
114. class(num2factor)
     
115. levels(num2factor)
     
116. table(num2factor)
     
117. 
     
118. ##################
     
119. # Code Snipped-11
     
120. ##################
     
121. 
     
122. # creating factor with given labels
     
123. num2factor <- cut(numvar,breaks=5,labels=c("lowest group","lower middle group", "middle group", "upper middle", "highest group"))
     
124. 
     
125. # displaying the data is tabular form
     
126. data.frame(table(num2factor))
     
127. 
     
128. # creating factor variable using conditional statement
     
129. num2factor <- factor(ifelse(numvar<=-1.37,1,ifelse(numvar<=-0.389,2,ifelse(numvar<=0.592,3,ifelse(numvar<=1.57,4,5)))),labels=c("(-2.35,-1.37]", "(-1.37,-0.389]", "(-0.389,0.592]", "(0.592,1.57]",   "(1.57,2.55]"))
     
130. 
     
131. # displaying data using table function
     
132. table(num2factor)
     
133. 
     
134. ##################
     
135. # Code Snipped-12
     
136. ##################
     
137. 
     
138. # creating date object using built in as.Date() function
     
139. as.Date("1970-01-01")
     
140. 
     
141. # looking at the internal value of date object
     
142. as.numeric(as.Date("1970-01-01"))
     
143. 
     
144. # Second January 1970 is showing number of elapsed day is 1.
     
145. as.Date("1970-01-02")
     
146. as.numeric(as.Date("1970-01-02"))
     
147. 
     
148. ##################
     
149. # Code Snipped-13
     
150. ##################
     
151. 
     
152. # creating date object specifying format of date
     
153. as.Date("Jan-01-1970",format="%b-%d-%Y")
     
154. 
     
155. ##################
     
156. # Code Snipped-14
     
157. ##################
     
158. 
     
159. # loading lubridate package
     
160. library(lubridate)
     
161. 
     
162. # creating date object using mdy() function
     
163. mdy("Jan-01-1970")
     
164. 
     
165. ##################
     
166. # Code Snipped-15
     
167. ##################
     
168. 
     
169. # creating heterogeneous date object
     
170. hetero_date <- c("second chapter due on 2013, august, 24", "first chapter submitted on 2013, 08, 18", "2013 aug 23")
     
171. 
     
172. # parsing the character date object and convert to valid date
     
173. ymd(hetero_date)
     
174. 
     
175. ##################
     
176. # Code Snipped-16
     
177. ##################
     
178. 
     
179. hetero_date <- c("second chapter due on 2013, august, 24", "first chapter submitted on 2013, 08, 18", "23 aug 2013")
     
180. ymd(hetero_date)
     
181. 
     
182. ##################
     
183. # Code Snipped-17
     
184. ##################
     
185. 
     
186. # Creating date object using based R functionality
     
187. date <- as.POSIXct("23-07-2013",format = "%d-%m-%Y", tz = "UTC")
     
188. date
     
189. 
     
190. # extracting month from the date object
     
191. as.numeric(format(date, "%m"))
     
192. 
     
193. # manipulating month by replacing month 7 to 8
     
194. date <- as.POSIXct(format(date,"%Y-8-%d"), tz = "UTC")
     
195. date
     
196. 
     
197. # The same operation is done using lubridate package
     
198. date <- dmy("23-07-2013")
     
199. date
     
200. 
     
201. month(date)
     
202. 
     
203. month(date) <- 8
     
204. date
     
205. 
     
206. ##################
     
207. # Code Snipped-18
     
208. ##################
     
209. 
     
210. # accessing system date and time
     
211. current_time <- now()
     
212. current_time
     
213. 
     
214. # changing time zone to "GMT"
     
215. current_time_gmt <- with_tz(current_time,"GMT")
     
216. current_time_gmt
     
217. 
     
218. # rounding the date to nearest day
     
219. round_date(current_time_gmt,"day")
     
220. 
     
221. # rounding the date to nearest month
     
222. round_date(current_time_gmt,"month")
     
223. 
     
224. # rounding date to nearest year
     
225. round_date(current_time_gmt,"year")
     
226. 
     
227. ##################
     
228. # Code Snipped-19
     
229. ##################
     
230. 
     
231. # creating a 10 element vector
     
232. num10 <- c(3,2,5,3,9,6,7,9,2,3)
     
233. # accessing 5th element
     
234. num10[5]
     
235. 
     
236. # checking whether there is any value of num10 object greater than 6
     
237. num10>6
     
238. 
     
239. # keeping only values greater than 6
     
240. num10[num10>6]
     
241. 
     
242. # use of negative subscript removes first element "3"
     
243. num10[-1]
     
244. 
     
245. ##################
     
246. # Code Snipped-20
     
247. ##################
     
248. 
     
249. # creating a data frame with 2 variables
     
250. data_2variable <- data.frame(x1=c(2,3,4,5,6),x2=c(5,6,7,8,1))
     
251. 
     
252. # accessing only first row
     
253. data_2variable[1,]
     
254. 
     
255. # accessing only first column
     
256. data_2variable[,1]
     
257. 
     
258. # accessing first row and first column
     
259. data_2variable[1,1]
     
260. 
     
261. ##################
     
262. # Code Snipped-21
     
263. ##################
     
264. 
     
265. list_obj<- list(dat=data_2variable,vec.obj=c(1,2,3))
     
266. list_obj
     
267. 
     
268. # accessing second element of the list_obj objects
     
269. list_obj[[2]]
     
270. list_obj[[2]][1]
     
271. 
     
272. # accessing dataset from the list object
     
273. list_obj$dat

复制代码

1. ##################
   
2. # Code Snipped-1
   
3. ##################
   
4. 
   
5. # notice that during split step a negative 5 is used within the code,
   
6. # this negative 5 has been used to discard fifth column of the iris data
   
7. # that contains "species" information and we do not need that column to calculate mean.
   
8. 
   
9. iris.set <- iris[iris$Species=="setosa",-5]
   
10. iris.versi <- iris[iris$Species=="versicolor",-5]
    
11. iris.virg <- iris[iris$Species=="virginica",-5]
    
12. 
    
13. # calculating mean for each piece ( The apply step)
    
14. mean.set <- colMeans(iris.set)
    
15. mean.versi <- colMeans(iris.versi)
    
16. mean.virg <- colMeans(iris.virg)
    
17. 
    
18. # combining the output (The combine step)
    
19. mean.iris <- rbind(mean.set,mean.versi,mean.virg)
    
20. 
    
21. # giving row names so that the output could be easily understood
    
22. rownames(mean.iris) <- c("setosa","versicolor","virginica")
    
23. 
    
24. ##################
    
25. # Code Snipped-2
    
26. ##################
    
27. 
    
28. # split-apply-combine using loop
    
29. # each iteration represents split
    
30. # mean calculation within each iteration represents apply step
    
31. # rbind command in each iteration represents combine step
    
32. 
    
33. mean.iris.loop <- NULL
    
34. for(species in unique(iris$Species))
    
35. {
    
36.   iris_sub <- iris[iris$Species==species,]
    
37.   column_means <- colMeans(iris_sub[,-5])
    
38.   mean.iris.loop <- rbind(mean.iris.loop,column_means)
    
39. }
    
40. 
    
41. # giving row names so that the output could be easily understood
    
42. rownames(mean.iris.loop) <- unique(iris$Species)
    
43. 
    
44. ##################
    
45. # Code Snipped-3
    
46. ##################
    
47. 
    
48. mean.iris.loop <- NULL
    
49. for(species in unique(iris$Species))
    
50. {
    
51.   iris_sub <- iris[iris$Species==species,]
    
52.   column_means <- colMeans(iris_sub[,-5])
    
53.   mean.iris.loop <- rbind(mean.iris.loop,column_means)
    
54. }
    
55. rownames(mean.iris.loop) <- unique(iris$Species)
    
56. 
    
57. mean.iris.loop
    
58. 
    
59. #The same mean calculation, but this time using the plyr package:
    
60. library(plyr)
    
61. ddply(iris,~Species,function(x) colMeans(x[,-
    
62. which(colnames(x)=="Species")]))
    
63. 
    
64. mean.iris.loop
    
65. 
    
66. ##################
    
67. # Code Snipped-4
    
68. ##################
    
69. 
    
70. # class of iris3 dataset is array
    
71. class(iris3)
    
72. # dimension of iris3 dataset
    
73. dim(iris3)
    
74. 
    
75. ##################
    
76. # Code Snipped-5
    
77. ##################
    
78. 
    
79. # Calculate column mean for each species and output will be data frame
    
80. iris_mean <- adply(iris3,3,colMeans)
    
81. 
    
82. class(iris_mean)
    
83. iris_mean
    
84. 
    
85. ##################
    
86. # Code Snipped-6
    
87. ##################
    
88. 
    
89. # again we will calculate the mean but this time output will be an array
    
90. iris_mean <- aaply(iris3,3,colMeans)
    
91. class(iris_mean)
    
92. iris_mean
    
93. 
    
94. # note that here the class is showing "matrix",
    
95. # since the output is a two dimensional array which represents matrix
    
96. 
    
97. # Now calculate mean again with output as list
    
98. iris_mean <- alply(iris3,3,colMeans)
    
99. class(iris_mean)
    
100. iris_mean
     
101. 
     
102. ##################
     
103. # Code Snipped-7
     
104. ##################
     
105. 
     
106. # converting 3 dimensional array to a 2 dimensional data frame
     
107. iris_dat <- adply(iris3, .margins=3)
     
108. class(iris_dat)
     
109. str(iris_dat)
     
110. 
     
111. ##################
     
112. # Code Snipped-8
     
113. ##################
     
114. 
     
115. # Function to calculate five number summary
     
116. fivenum.summary <- function(x)
     
117. {
     
118.   results <-data.frame(min=apply(x,2,min),
     
119.   mean=apply(x,2,mean),
     
120.   median=apply(x,2,median),
     
121.   max=apply(x,2,max),
     
122.   sd=apply(x,2,sd))
     
123.   return(results)
     
124. }
     
125. 
     
126. #To calculate the summaries for the five numbers using a for loop with default R is as shown:
     
127. # initialize the output list object
     
128. all_stats <- list()
     
129. 
     
130. # the for loop will run for each species
     
131. for(i in 1:dim(iris3)[3])
     
132. {
     
133.   sub_data <- iris3[,,i]
     
134.   all_stat_species <- fivenum.summary(sub_data)
     
135.   all_stats[[i]] <-  all_stat_species
     
136. }
     
137. 
     
138. # class of the output object
     
139. class(all_stats)
     
140. all_stats
     
141. 
     
142. # Let's calculate the same statistics, but this time using the adply() function from the plyr package:
     
143. all_stats <- alply(iris3,3,fivenum.summary)
     
144. 
     
145. class(all_stats)
     
146. all_stats
     
147. 
     
148. ##################
     
149. # Code Snipped-9
     
150. ##################
     
151. 
     
152. # define parameter set
     
153. parameter.dat <- data.frame(n=c(25,50,100,200,400),mean=c(0,2,3.5,2.5,0.1),sd=c(1,1.5,2,5,2))
     
154. 
     
155. # displaying parameter set
     
156. 
     
157. parameter.dat
     
158. 
     
159. # random normal variate generate using base R
     
160. # set seed to make the example reproducible
     
161. set.seed(12345)
     
162. 
     
163. # initialize blank list object to store the generated variable
     
164. dat <- list()
     
165. for(i in 1:nrow(parameter.dat))
     
166. {
     
167. dat[[i]] <- rnorm(n=parameter.dat[i,1],
     
168.              mean=parameter.dat[i,2],sd=parameter.dat[i,3])
     
169. }
     
170. 
     
171. # estimating mean from the newly generated data
     
172. estmean <- lapply(dat,mean)
     
173. estmean
     
174. 
     
175. # Performing same task as above but this time use plyr package
     
176. 
     
177. dat_plyr <- mlply(parameter.dat,rnorm)
     
178. estmean_plyr <- llply(dat_plyr,mean)
     
179. estmean_plyr

复制代码

1. ##################
   
2. # Code Snipped-1
   
3. ##################
   
4. 
   
5. # Example of typical two dimensional data
   
6. 
   
7. # A demo dataset "students" with typical layout. This data contains
   
8. # two students' exam score of "math", "literature" and "language" in
   
9. # different term exam.
   
10. students <- data.frame(sid=c(1,1,2,2),
    
11.                        exmterm=c(1,2,1,2),
    
12.                        math=c(50,65,75,69),
    
13.                        literature=c(40,45,55,59),
    
14.                        language=c(70,80,75,78))
    
15. students
    
16. 
    
17. ##################
    
18. # Code Snipped-2
    
19. ##################
    
20. 
    
21. library(reshape)
    
22. # Example of molten data
    
23. molten_students <- melt.data.frame(students,id.vars=c("sid","exmterm"))
    
24. 
    
25. ##################
    
26. # Code Snipped-3
    
27. ##################
    
28. 
    
29. # Reshaping dataset using reshape function
    
30. wide_students <- reshape(students,direction="wide",idvar="sid",timevar="exmterm")
    
31. wide_students
    
32. 
    
33. # Now again reshape to long format
    
34. long_students <- reshape(wide_students,direction="long",idvar="id")
    
35. long_students
    
36. 
    
37. ##################
    
38. # Code Snipped-4
    
39. ##################
    
40. 
    
41. # original data
    
42. students
    
43. 
    
44. # Melting by specifying both id and measured variables
    
45. melt(students,id=c("sid","exmterm"), measured=c("math","literature","language"))
    
46. 
    
47. # Melting by specifying only id variables
    
48. melt(students,id=c("sid","exmterm"))
    
49. 
    
50. ##################
    
51. # Code Snipped-5
    
52. ##################
    
53. 
    
54. # Melting students data
    
55. molten_students <- melt(students,id.vars=c("sid","exmterm"))
    
56. molten_students
    
57. 
    
58. # return back to original data
    
59. cast(molten_students,sid+exmterm~variable)
    
60. 
    
61. # Now the same operation but specifying only row variable.
    
62. cast(molten_students,...~variable)
    
63. 
    
64. # We now rearrange the data where sid is now separate column for each student
    
65. cast(molten_students,...~sid)
    
66. 
    
67. # Again rearranging the data where exmterm is now separate column for each term
    
68. cast(molten_students,...~exmterm)

复制代码

1. ##################
   
2. # Code Snipped-1
   
3. ##################
   
4. 
   
5. # Trying to create a vector of zero with length 2^32-1. Note that the RAM
   
6. # of the computer we are generating this example is 8GB with 64-bit Windows-7
   
7. # Professional edition.Processor core i5.
   
8. 
   
9. x <- rep(0, 2^31-1)
   
10. 2^31
    
11. 
    
12. # If we try to assign a vector of length greater than maximum addressable
    
13. # length then that will produce NA
    
14. 
    
15. as.integer(2^31)
    
16. 
    
17. ##################
    
18. # Code Snipped-2
    
19. ##################
    
20. 
    
21. # calling ODBC library into R
    
22. library(RODBC)
    
23. 
    
24. # creating connection with the database using odbc package and the connection
    
25. # we created earlier.
    
26. 
    
27. xldb<- odbcConnect("xlopen")
    
28. 
    
29. # In the odbcConnect() function the minimum argument required
    
30. # is the ODBC connection string.
    
31. 
    
32. # Now the connection created, using that connection we will import data
    
33. 
    
34. xldata<- sqlFetch(xldb, "CSVanscombe")
    
35. 
    
36. # Note here that "CSVanscombe"is the Excel worksheet name.
    
37. 
    
38. odbcClose(xldb) # closing the database connection
    
39. 
    
40. ##################
    
41. # Code Snipped-3
    
42. ##################
    
43. 
    
44. # calling odbc library
    
45. library(RODBC)
    
46. 
    
47. # connecting with database
    
48. access_con<- odbcConnect("accessdata")
    
49. 
    
50. # import separate table as separate R data frame
    
51. coverage_page<- sqlFetch(access_con, "coverpage")
    
52. ques1 <- sqlFetch(access_con, "questionnaire1")
    
53. ques2 <- sqlFetch(access_con, "questionnaire2")
    
54. 
    
55. odbcClose(access_con) # closing the database connection
    
56. ##################
    
57. # Code Snipped-4  for filehash package
    
58. ##################
    
59. 
    
60. library(filehash)
    
61. dbCreate("exampledb")
    
62. filehash_db<- dbInit("exampledb")
    
63. 
    
64. dbInsert(filehash_db, "xx", rnorm(50))
    
65. value<- dbFetch(filehash_db, "xx")
    
66. summary(value)
    
67. 
    
68. dbInsert(filehash_db, "y", 4709)
    
69. dbDelete(filehash_db, "xx")
    
70. dbList(filehash_db)
    
71. dbExists(filehash_db, "xx")
    
72. 
    
73. filehash_db$x<- runif(100)
    
74. summary(filehash_db$x)
    
75. summary(filehash_db[["x"]])
    
76. filehash_db$y<- rnorm(100, 2)
    
77. dbList(filehash_db)
    
78. 
    
79. # To run the following line make sure the working directory is set properly.
    
80. # The working directory should be the folder where the file "anscombe.txt" is stored
    
81. 
    
82. dumpDF(read.table("anscombe.txt", header=T), dbName="massivedata")
    
83. massive_environment<- db2env(db="massivedata")
    
84. 
    
85. fit<- with(massive_environment, lm(Y1~X1))
    
86. with(massive_environment, summary(Y1))
    
87. with(massive_environment, Y1[1] <- 99)
    
88. 
    
89. ##################
    
90. # Code Snipped-5  for ff package
    
91. ##################
    
92. library(ff)
    
93. file1 <- ff(filename="file1", length=10,vmode="double")
    
94. str(file1)
    
95. 
    
96. # calling rivers data
    
97. data(rivers)
    
98. file1[1:10] <- rivers[1:10]
    
99. 
    
100. # Note that here file1 is an ff object whereas
     
101. # file1[...] returns default R vector
     
102. str(file1)
     
103. 
     
104. # We can perform sampling if required on the ff objects:
     
105. # set seed to reproduce the example
     
106. set.seed(1337)
     
107. sample(file1,5,replace=FALSE)
     
108. 
     
109. gc()
     
110. 
     
111. ##################
     
112. # Code Snipped-6  for sqldf package
     
113. ##################
     
114. 
     
115. # Selecting the rows from iris dataset where sepal length > 2.5
     
116. # and store that in subiris data frame
     
117. 
     
118. library(sqldf)
     
119. subiris<- sqldf("select * from iris where Sepal_Width> 3")
     
120. head(subiris)
     
121. nrow(subiris)
     
122. 
     
123. subiris2<- sqldf("select Sepal_Length,Petal_Length,Species from iris where Petal_Length> 1.4")
     
124. nrow(subiris2)
     
125. 
     
126. # Before running the following line, make sure the working directory is set properly
     
127. # import only Sepal width and Petal width along with species information where Petal width is greater than 0.4
     
128. iriscsv<-read.csv.sql("iris.csv",sql="select Sepal_Width,Petal_Width,Species from file where Petal_Width>0.4")
     
129. head(iriscsv)
     
130. 
     
131. # do not use underscore as within variable name it will give error, here is the example
     
132. iriscsv<-read.csv.sql("iris.csv",sql="select Sepal.Width,Petal.Width,Species from file where Petal.Width>0.4")
     
133. 
     
134. # we can draw a random sample of size 10 from iris data that are stored in iris.csv file.
     
135. iris_sample<- read.csv.sql("iris.csv",sql="select * from file order by random(*) limit 10")
     
136. iris_sample
     
137. 
     
138. # Calculate group wise mean from iris data
     
139. iris_avg<-sqldf("select Species, avg(Sepal_Length),avg(Sepal_Width),avg(Petal_Length),avg(Petal_Width) from iris group by Species")
     
140. 
     
141. colnames(iris_avg) <- c("Species","Sepal_L","Sepal_W","Petal_L","Petal_W")
     
142. iris_avg
     
143. 
     
144. # The base R counterpart to perform same operation is
     
145. aggregate(iris[,-5],list(iris$Species),mean)

复制代码

kankan

书在哪里？

have a look

ReneeBK 发表于 2014-11-11 06:31
https://www.packtpub.com/big-data-and-business-intelligence/data-manipulation-r
Data Manipulation w ...

good