比较R中连续行的值

Question

我有一个数据表，对于每个datatable$Ppt和每个datatable$nitem，当datatable$Region中有"fffword“时，我需要提取"fffword”的值，并将其与以下"word“的值进行比较。如果这两个值是相同的，那么在datatable$Output中我需要一个0的值，如果这两个值是不同的，我需要datatable$Output中的值1。

我试过：

 datatable %>% group_by(Ppt, nitem) %>%
   mutate(Output = ifelse(as.numeric(gsub("fffword([0-9]+).*","\\1",Region) == lag(as.numeric(gsub("word([0-9]+).*","\\1",Region)), 0L,ifelse(as.numeric(gsub("fffword([0-9]+).*","\\1",Region) != lag(as.numeric(gsub("word([0-9]+).*","\\1",Region)), 1L)

但不起作用。

 #Ppt      Region            nitem      Output
 #1        "fffword8"        93         0 (current ffword n=8, following word n=8)
 #1        "word8"           93         0 (previous ffword n=8, current word n=8)
 #1        "fffword9"        93         1 (current ffword n=9, no following word for this ppt and this nitem)
 #1        "word2"           122        1 (no previous fffword for this ppt and this nitem and this n Region)
 #1        "fffword3"        122        0 (current ffword n=3, following word n=3)
 #1        "word3"           122        0 (previous ffword n=3, current word n=3)
 #1        "word6"           122        1 (no previous fffword for this ppt and this nitem and this n Region)
 #1        "fffword7"        122        0
 #1        "word7"           122        0
 #1        "fffword8"        122        0
 #1        "word8"           122        0
 #54       "fffword8"        4          0
 #54       "word8"           4          0
 #54       "fffword9"        4          1
 #54       "word2"           4          1
 #54       "fffword2"        10         0
 #54       "word4"           10         1
 #54       "word6"           10         1
 #54       "fffword23"       10         0
 #54       "word23"          10         0
 #54       "fffword24"       5          0
 #54       "word24"          5          0

Answer 1

使用嵌套的ifelse和dplyr进行一次尝试可以提供您想要的结果。办法是：

如果当前行的区域为fffword，则应将数字与next(lead)行匹配，但当当前行的区域为word时，则应将数字与前一行(lag)行进行比较。

如果next或previous行不可用，则应将Output视为1。

来自区域的digit刚刚被比较为character格式。在进行比较之前，没有明显的理由在numeric中转换这些内容。

library(dplyr)
datatable  <- read.table(text = 'Ppt      Region     nitem      
1        "fffword8"        93         
1        "word8"           93          
1        "fffword9"        93          
1        "word2"           122         
1        "fffword3"        122         
1        "word3"           122         
1        "word6"           122         
1        "fffword7"        122        
1        "word7"           122        
1        "fffword8"        122        
1        "word8"           122        
54       "fffword8"        4          
54       "word8"           4          
54       "fffword9"        4          
54       "word2"           4          
54       "fffword2"        10         
54       "word4"           10         
54       "word6"           10         
54       "fffword23"       10         
54       "word23"          10         
54       "fffword24"       5          
#54       "word24"          5         ', header = T, stringsAsFactors = F)


datatable %>% group_by(Ppt, nitem) %>%
  mutate(Output = ifelse(grepl("^fffword",Region),
                          ifelse(is.na(lead(Region)),1L,
                                 ifelse((gsub("fffword([0-9]+).*","\\1",Region)) == (gsub("word([0-9]+).*","\\1",lead(Region))), 0L,1L)),
                          ifelse(is.na(lag(Region)),1L,
                                 ifelse((gsub("word([0-9]+).*","\\1",Region)) == (gsub("fffword([0-9]+).*","\\1",lag(Region))), 0L,1L))
                          )) %>% as.data.frame()


#Result
   Ppt    Region nitem Output
1    1  fffword8    93      0
2    1     word8    93      0
3    1  fffword9    93      1
4    1     word2   122      1
5    1  fffword3   122      0
6    1     word3   122      0
7    1     word6   122      1
8    1  fffword7   122      0
9    1     word7   122      0
10   1  fffword8   122      0
11   1     word8   122      0
12  54  fffword8     4      0
13  54     word8     4      0
14  54  fffword9     4      1
15  54     word2     4      1
16  54  fffword2    10      1
17  54     word4    10      1
18  54     word6    10      1
19  54 fffword23    10      0
20  54    word23    10      0
21  54 fffword24     5      1