dplyr::summarize() 返回平均值的 NA

Question

我想创建一个数据集，使用有向图可视化每日最低、最高和平均温度。我有一个时间序列数据集，其中一列中包含日期（日期时间戳），后续列中包含水质（温度、pH 等）变量。数据每隔 15 分钟报告一次。数据集中存在传感器发生故障的 NA。

数据帧 ldq1 的标头如下所示：

 datetimestamp temp 
1  2007-01-01 00:00:00 16.4  
2  2007-01-01 00:15:00 16.4  
3  2007-01-01 00:30:00 16.4  
4  2007-01-01 00:45:00 16.4  
5  2007-01-01 01:00:00 16.4  
6  2007-01-01 01:15:00 16.4  
7  2007-01-01 01:30:00 16.3  
8  2007-01-01 01:45:00 16.2  
9  2007-01-01 02:00:00 16.3  

The datetimestamp class:
class(ldq1$datetimestamp)
[1] "POSIXct" "POSIXt" 

温度是双倍。

我正在总结每日的最大值、分钟和平均值，如下所示：

l<-ldq1 %>%
+   group_by(date = as.Date(datetimestamp)) %>%
+   summarise(t_min = min(temp, na.rm = TRUE),
+             t_max = max(temp, na.rm = TRUE),
+             t_mean = mean(temp, na.rm = TRUE), .groups = 'drop')

但是输出看起来像这样：

Warning message:
There were 6705 warnings in `summarise()`.
The first warning was:
ℹ In argument: `t_min = min(temp, na.rm = TRUE)`.
ℹ In group 187: `date = 2007-07-06`.
Caused by warning in `min()`:
! no non-missing arguments, returning NA

l

# A tibble: 5,845 × 4
   date       t_min t_max t_mean
   <date>     <chr> <chr>  <dbl>
 1 2007-01-01 15.7  17.1      NA
 2 2007-01-02 15.2  16.7      NA
 3 2007-01-03 14.7  15.7      NA
 4 2007-01-04 15.3  16.1      NA
 5 2007-01-05 15.6  16.9      NA
 6 2007-01-06 15.9  17.3      NA
 7 2007-01-07 16.3  17.8      NA
 8 2007-01-08 16.7  17.8      NA
 9 2007-01-09 16    17.5      NA
10 2007-01-10 14.7  16.7      NA

我哪里出错了？

Answer 1

正如 @Dave2e 和 @Jon Spring 已经提到的，问题是你的

temp

列被编码为字符变量而不是数字。因此，您只需添加一个步骤即可使用

mutate

和

across

进行此转换。

ldq1 %>%
  # Transform temp from char to numeric
  mutate(across(temp, ~as.numeric(.x))) %>%
  group_by(date = as.Date(datetimestamp)) %>%
  summarise(t_min = min(temp, na.rm = TRUE),
            t_max = max(temp, na.rm = TRUE),
            t_mean = mean(temp, na.rm = TRUE), .groups = 'drop')

# A tibble: 1 × 4
#  date       t_min t_max t_mean
#  <date>     <dbl> <dbl>  <dbl>
#1 2007-01-01  16.2  16.4   16.4