dplyr count observations per day

问题:

I have the following data

Name         Date                                   Message
Ted Foe      2011-06-10T05:06:30+0000               I love this product
Sina Fall    2011-06-10T05:07:33+0000               Not my type of product
Steve Hoe    2011-06-11T05:06:30+0000               Great Discussion! Thanks
Selda Dee    2011-06-13T05:12:30+0000               Seen elsewhere
Steven Hoe   2011-06-13T03:17:31+0000               Where?
Selda Dee    2011-06-13T05:17:56+0000               Tinder

I want to aggregate by days so that I end up with a time series like this

Date            Number of Posts
2011-06-10      2
2011-06-11      1
2011-06-12      0
2011-06-13      3

I already tried the following

summary_df <- df %>% group_by(Date) %>% summarise(comments = count(message))

But this is not working. Any quick dplyr based solution would be great.

Thanks for the help!

Cheers, Raoul

回答1:

Grouped by the 'Date' column after converting to Date class, we get the number of rows (n()) with summarise. If we need the 'Date' elements that are missing in the original dataset, create a new dataset with the sequence of minimum to maximum 'Date' and do a left_join

df1 <- df %>%
          group_by(Date = as.Date(Date)) %>%
          summarise(comments = n())
expand.grid(Date = seq(min(df1$Date), max(df1$Date), by = '1 day')) %>%
         left_join(., df1)