scale_fill_manual not respecting to the order of values and labels-CodePudding

For the code below, I try to use scale_fill_manual('trend', values = c('red', 'blue', 'gray'), labels = c('up', 'down', 'keep')) to set bar color for each month with red for up, blue for down and gray for keep:

df <- structure(list(month = structure(c(1L, 1L, 2L, 2L, 3L, 3L, 4L, 
4L, 5L, 5L, 6L, 6L, 7L, 7L, 8L, 8L, 9L, 9L, 9L, 10L, 10L, 11L, 
11L, 12L, 12L), .Label = c("1", "2", "3", "4", "5", "6", "7", 
"8", "9", "10", "11", "12"), class = "factor"), trend = c("down", 
"up", "down", "up", "down", "up", "down", "up", "down", "up", 
"down", "up", "down", "up", "down", "up", "down", "keep", "up", 
"down", "up", "down", "up", "down", "up"), numbers = c(10L, 16L, 
16L, 11L, 15L, 12L, 14L, 13L, 15L, 12L, 15L, 12L, 14L, 13L, 16L, 
11L, 14L, 1L, 12L, 15L, 12L, 14L, 13L, 15L, 12L)), class = "data.frame", row.names = c(NA, 
-25L))

ggplot(df, aes(x = month, y = numbers))  
  geom_bar(aes(fill = trend), alpha=0.8,
           stat='identity', position = 'dodge', width = 0.7)  
  scale_fill_manual('trend', values = c('red', 'blue', 'gray'),
                    labels = c('up', 'down', 'keep'))  
  geom_text(aes(x = month, y = numbers, label=numbers, group=trend), position = position_dodge(width=0.9), vjust=-0.25)  
  labs(x='month', 
       y='numbers')  
  theme(
    legend.position='bottom', 
    legend.title = element_blank()
  )

Now the problem we can find in this figure is: red for down, blue for keep and gray for up, which is different from the requested.

If we comment scale_fill_manual(...) part and rerun the code, the default order correspondence is correct:

So I'm wondering why I get this issue and what's the correct method to set customized colors for this specific data (only one keep entry for September group)? Thanks.

Data:

   month trend numbers
1      1  down      10
2      1    up      16
3      2  down      16
4      2    up      11
5      3  down      15
6      3    up      12
7      4  down      14
8      4    up      13
9      5  down      15
10     5    up      12
11     6  down      15
12     6    up      12
13     7  down      14
14     7    up      13
15     8  down      16
16     8    up      11
17     9  down      14
18     9  keep       1
19     9    up      12
20    10  down      15
21    10    up      12
22    11  down      14
23    11    up      13
24    12  down      15
25    12    up      12

CodePudding user response：

You can specify values as named character vector in scale_fill_manual:

ggplot(df, aes(x = month, y = numbers))  
  geom_bar(aes(fill = trend), alpha=0.8,
           stat='identity', position = 'dodge', width = 0.7)  
  scale_fill_manual('trend',
                    values = c(up = 'red', down = 'blue', keep = 'gray'))  
  geom_text(aes(x = month, y = numbers, label=numbers, group=trend), position = position_dodge(width=0.9), vjust=-0.25)  
  labs(x='month', 
       y='numbers')  
  theme(
    legend.position='bottom', 
    legend.title = element_blank()
  )