dplyr-summarise 数据分组进行统计描述

summarise()在dplyr中函数用于对数据进行统计描述。

    maingroup subgroup        value      norm
1       AAA      Y.1  0.152138053 10.717161
2       AAA      Y.1  0.129267562  8.940263
3       BBB      X.1 -0.360278960  9.723177
...

#变量类型查看
> df %>% str()
'data.frame':	3000 obs. of  4 variables:
 $ maingroup: chr  "AAA" "AAA" "AAA" "BBB" ...
 $ subgroup: chr  "Y.1" "Y.1" "X.1" "X.1" ...
 $ value   : num  0.152 0.129 0.892 -0.36 -0.224 ...
 $ norm    : num  10.72 8.94 9.67 9.72 10.59 ...

1、统计所有列的空值数

df %>% summarise_all(  ~ sum(is.na(.)) )
>...
  maingroup subgroup value norm
1         0        0     0    0

2、分组统计每组的记录数n()

df %>% group_by(species) %>% summarise(  n = n() )
>...
# A tibble: 6 x 3
# Groups:   maingroup [3]
  maingroup subgroup     n
  <chr>     <chr>    <int>
1 AAA       X.1        494
2 AAA       Y.1        506
3 BB        X.1        504
4 BB        Y.1        496
5 CCC       X.1        522
6 CCC       Y.1        478

###3、 除开分组的列向量，对其它所有列进行分组求和

df %>% group_by(maingroup ,subgroup) %>% summarise(across(everything(), sum, na.rm = TRUE))
>...
# A tibble: 6 x 4
# Groups:   rowname [3]
 maingroup subgroup  value  norm
  <chr>   <chr>     <dbl> <dbl>
1 AAA     X.1      -4.57  4969.
2 AAA     Y.1      10.3   5063.
3 BB      X.1       7.11  4906.
4 BB      Y.1      -6.47  5074.
5 CCC     X.1       0.411 4754.
6 CCC     Y.1      -7.94  5239.

###4、对所有列(不包括用于分组的列向量)进行分组求和，求均值

df %>% group_by(maingroup,subgroup ) %>% summarise(across(everything(), list(mean = mean, sum = sum),na.rm = TRUE))
>...
# A tibble: 6 x 6
# Groups:   maingroup [3]
  maingroup subgroup value_mean value_sum norm_mean norm_sum
  <chr>     <chr>         <dbl>     <dbl>     <dbl>    <dbl>
1 AAA       X.1        -0.00708     -3.58     10.0     5059.
2 AAA       Y.1        -0.0312     -15.4      10.1     4984.
3 BB        X.1         0.0358      18.4       9.94    5108.
4 BB        Y.1        -0.0328     -16.0      10.1     4894.
5 CCC       X.1         0.0270      13.3      10.0     4945.
6 CCC       Y.1        -0.00361     -1.83      9.97    5054.

5、求取列向量的分位数

df %>% summarise(across(c(value ,norm), ~quantile(.,na.rm = TRUE)) ) 
>...
         value      norm
1 -0.999550791  6.331832
2 -0.484380580  9.351605
3 -0.009638345 10.029124
4  0.487067557 10.691726
5  0.999956779 13.868918

6、 对指定类型的列进行统计

df %>%  summarise( across(where(is.numeric) & !norm, ~scale(.x)) ) %>% colSums()
>...
value norm
0 0