expand functions added

[Quiri]

Hello Hadley,

i'm a Data Analyst from Germany and love working with your packages, especially dplyr. While using the group_by %>% summarize chain, I discovered, that there are only groups made out of existing data. But if I want to group by date, for example, and I don't have data for a specific date, then its does not appear in the grouped result, so I have a "hole", but I would like to have

Date	n()
2014-01-01	5
2014-01-02	0 instead of no data at all
2014-01-03	7

and so on...

I didn't find a solution, so I build a function for it based on expand.grid(). I thought it would be a nice add to tidyr, so take a look at it, maybe you like it (or at least the idea of expanding data).

You can expand from min:max and by unique ids (which is convenient if you grouped by many variables, like date and gender), and you can select as many expandable columns as you wish.
It works great with the dplyr pipe and I use it a lot.

Feel free to contact me.

[Quiri]

Here is an example how expand could be used in a dplyr chain:
Assume we look at payment data and want to have a table with total payers up to a respective day.

firstpay <- data%>%
  group_by(UserId)%>% 
  filter(paytime == min(paytime))%>%           # Filter the data with the first payment per User
  group_by(date)%>%
  summarize(first_time_payer = n())%>%     # Calculate the amount of first time payers per day
  ungroup%>%                                             # Ungroup for cumsum operator
  expand("date")%>%                                   # Expand by date, since there are days without first time payers
  mutate(total_payer = cumsum(first_time_payer))  # calculate the cumulated amount payers for every date point

[hadley]

Thanks! I liked the idea, but ended up going in a somewhat different direction for the implementation.

[Quiri]

Awesome! That's actually what I was hoping for ;)