时间戳

首页 > 解决方案 > 跨多列在行内比较删除不匹配并创建新行

r - 跨多列在行内比较删除不匹配并创建新行

问题描述

我正在尝试计算他相同的地址并按行分组。我相当接近，但在特定地址之间的列之间存在细微差别。目的是从行中删除任何不匹配的地址，并将它们作为新行添加到 df. 街道号或街区号之间通常存在差异。我已经从代码广告中提取了这些数字，我试图找到那些不匹配的数字，删除它们并创建一个新行并适当地更改计数。计数更改可以在之后进行，只需检查行中的非缺失。

该数据集实际上有 5000 行，一行最多 50 个建筑物。这是一个示例。

 df<-data.frame(bldg1 = c("26 this street, big district","block8, fancy estate, small district", "11 normal lane, district"),
                bldg2 = c("27 this street, big district","block8, fancy estate, small district", "11 normal lane, district"),
                bldg3 = c("26 this street, big district","block6, fancy estate, small district", "11 normal lane, district"),
                bldg4 = c("26 this street, big district","block8, fancy estate, small district", NA),
                bldg5 = c("26 this street, big district","block6, fancy estate, small district", "11 normal lane, district"),
                bldg1strnum = c("26",NA, "11"),
                bldg2strnum = c("27",NA, "11"),
                bldg3strnum = c("26",NA, "11"),
                bldg4strnum = c("26",NA, "11"),
                bldg5strnum = c("26",NA, "11"),
                bldg1blck = c(NA,"8", NA),
                bldg2blck = c(NA,"8", NA),
                bldg3blck = c(NA,"6", NA),
                bldg4blck = c(NA,"8", NA),
                bldg5blck = c(NA,"6", NA),
               count = (5,5,4))

我正在考虑使用dplyrand acrosswithlength(unique)但不知道如何正确运行它，尤其是如何将mutate其转换为新行的长格式。

我喜欢的结果如下所示。（突变后不需要街道号码和名称

 df<-data.frame(bldg1 = c("26 this street, big district","block8, fancy estate, small district", "11 normal lane, district", "27 this street, big district","block6, fancy estate, small district"),
               bldg2 = c(NA, "block8, fancy estate, small district", "11 normal lane, district",NA,"block6, fancy estate, small district"),
               bldg3 = c("26 this street, big district",NA, "11 normal lane, district", NA, NA),
               bldg4 = c("26 this street, big district","block8, fancy estate, small district", NA,NA,NA),
               bldg5 = c("26 this street, big district",NA, "11 normal lane, district",NA,NA),
               count = ("4","3","4","1","2"))

标签： rdplyracross

解决方案

这是你想要的：

df %>% 
  select(bldg1, bldg2, bldg3, bldg4, bldg5) %>% 
  pivot_longer(
    cols = everything()
  ) %>% 
  arrange(value) %>% 
  add_count(value)

输出：

   name  value                                    n
   <chr> <chr>                                <int>
 1 bldg1 11 normal lane, district                 4
 2 bldg2 11 normal lane, district                 4
 3 bldg3 11 normal lane, district                 4
 4 bldg5 11 normal lane, district                 4
 5 bldg1 26 this street, big district             4
 6 bldg3 26 this street, big district             4
 7 bldg4 26 this street, big district             4
 8 bldg5 26 this street, big district             4
 9 bldg2 27 this street, big district             1
10 bldg3 block6, fancy estate, small district     2
11 bldg5 block6, fancy estate, small district     2
12 bldg1 block8, fancy estate, small district     3
13 bldg2 block8, fancy estate, small district     3
14 bldg4 block8, fancy estate, small district     3
15 bldg4 NA                                       1

推荐阅读