R：堆叠多列

Question

I'm trying to reorganize a data set I pulled online. The table online is formatted so that there are two tables side by side using both the same column name. (Date, Price | Date, Price).

>head(data)
                    Date  Price             Date.1 Price.1
    1  December 17, 2016  1,546bp December 7, 2016 1,720bp
    2  December 16, 2016  1,628bp December 6, 2016 1,638bp
    3  December 15, 2016  1,714bp December 5, 2016 1,560bp
    4  December 14, 2016  1,805bp December 4, 2016 1,511bp
    5  December 13, 2016  1,900bp December 3, 2016 1,440bp

So what I'm trying to do is stack the 2 variables on the right underneath the 2 variables on the left to make it in chronological order.

I've tried the stack function and the reshape2 package in R and it won't work because the data isn't all numeric. I've tried other methods of creating data frames of just the two date variables and then stacking them within the new data frame but it still doesn't work. Regardless, below is the code I used to try to stack the variables.

> melt(newtable12, id.vars=c('Date', 'Date.1'), variable.name='DD')
    Date                  Date.1              DD    value
    1  December 17, 2016  December 7, 2016    Price 1,546bp
    2  December 16, 2016  December 6, 2016    Price 1,628bp
    3  December 15, 2016  December 5, 2016    Price 1,714bp
    4  December 14, 2016  December 4, 2016    Price 1,805bp
    5  December 13, 2016  December 3, 2016    Price 1,900bp
    6  December 12, 2016  December 2, 2016    Price 2,000bp
    7  December 11, 2016  December 1, 2016    Price 1,926bp
    8  December 10, 2016 November 30, 2016    Price 1,834bp
    9  December 9, 2016  November 29, 2016    Price 1,746bp
   10  December 8, 2016  November 28, 2016    Price 1,771bp

Could anyone provide any insight on correctly stacking the date/date.1 and price/price.1 columns? I'm somewhat new to R if that helps. Thank you in advance.

Answer 1

I think this should do the trick - the last line may not be necessary, as your data appears to already be in the correct order.

data[nrow(data) + 1:nrow(data), 1:2] <- data[, 3:4]
data <- data[, 1:2]                                  # Remove excess columns

If any re-ordering is still necessary, then if the 'date' column is formatted as dates, rather than strings, you could try:

data <- data[order(data$Date, decreasing=T), ]       # Order by date

Otherwise, if the 'date' column is made up of characters, you could adapt something like this (although it's far from a perfect solution):

# Extract date information
data$month <- unlist(strsplit(data$Date, " "))[c(T,F,F)]
data$year <- unlist(strsplit(data$Date, " "))[c(F,F,T)]
data$day <- unlist(strsplit(data$Date, " "))[c(F,T,F)]
data$month <- factor(data$month, levels=month.name)
data$day <- sapply(data$day, function(x) substr(x, 1, nchar(x)-1))

# Order the dataframe
data <- data[order(year, month, day, decreasing=TRUE)