Question

I have a dataframe like so:

item   price source   day
book   5     shop     2012-01-01
car    100   shop     2012-01-01
desk   10    shop     2012-01-01
book   4     internet 2012-01-01
car    99    internet 2012-01-01
desk   9     internet 2012-01-01
book   6     shop     2012-01-02
car    101   shop     2012-01-02
desk   11    shop     2012-01-02
book   4     internet 2012-01-02
car    100   internet 2012-01-02
desk   10    internet 2012-01-02

I would like, per day, each price per item, like:

item price.shop price.internet day
book 5          4              2012-01-01
car  100        99             2012-01-01
desk 10         9              2012-01-01
book 6          5              2012-01-02
car  101        100            2012-01-02
desk 11         10             2012-01-02

I have been trying my best with

melt‘ing and then

cast(df, item ~ source + variable)

But can’t get the desired result. I think the issue is becase i have in essence two “per thing” conditions (per day, per source) and so dont have the correct equation.

Thanks for any help

Answer 1

Not a plyr.reshape2 answer but using base:reshape():

> reshape(dat, idvar=c("item", "day"),  timevar="source", direction="wide")
  item        day price.shop price.internet
1 book 2012-01-01          5              4
2  car 2012-01-01        100             99
3 desk 2012-01-01         10              9
7 book 2012-01-02          6              4
8  car 2012-01-02        101            100
9 desk 2012-01-02         11             10

I suppose one could also could have tried a merge(subset_shop, subset_internet, 1:2, all=TRUE) strategy.