generate seq of quarter date in R - r

Question

I am new to R and am I have a data frame that looks something like this.
Date A B
1990 Q1 2 3
Q2 4 2
Q3 7 6
Q4 5 3
1991 Q1 7 6
Q2 1 8
Q3 7 6
Q4 9 2
1992 Q1 1 7
Q2 4 6
Q3 1 3
Q4 5 8
...
The column stretches all the way to the end of the row and both the start date and the end date is not fixed as the data is constantly updated. I would like to format the date column into a date class and achieve something like this:
Date A B
1990 Q1 2 3
1990 Q2 4 2
1990 Q3 7 6
1990 Q4 5 3
1991 Q1 7 6
1991 Q2 1 8
1991 Q3 7 6
1991 Q4 9 2
1992 Q1 1 7
1992 Q2 4 6
1992 Q3 1 3
1992 Q4 5 8
...
I thought of recreating a new column of dates on the left and use the first date provided by the data (i.e. '1990 Q1') as the starting date and the length based on the number of rows. Was looking at using seq. and as.yearqtr commands but can't seem to work out a proper code for it. Anyone knows of a better way to do this?

Answer 1

To use the yearqtr function from the zoo package to create a year-quarter time series, you can first split the df$Date values into year and quarter strings, use na.locf, also from the zoo package, to fill in missing values of year with the value from the previous row, and then transform to a zoo time series with year quarter dates. Code would look like
library(zoo)
# split Date into year and quarter strings
tmp <- t(sapply(strsplit((df$Date), " "), function(x) if(length(x)==1) c(NA, x) else x))
# use na.locf to replace NA with previous year
tmp <- paste(na.locf(tmp[,1]), tmp[,2])
# transform df into a zoo time series object with yearqtr dates
df_zoo <- zoo(df[,-1], order.by = as.yearqtr(tmp))

Answer 2

We could do this in base R. Create a grouping variable using grep and cumsum, extract the numeric substring from 'Date', replace the '' values with the year values using ave, and then paste it with the quarter substring extracted using sub.
df$Date <- paste(ave(sub("\\s*Q.", "", df$Date),
cumsum(grepl("^\\d+", df$Date)), FUN = function(x) x[nzchar(x)]),
sub("^\\d+\\s+", "", df$Date))
df$Date
#[1] "1990 Q1" "1990 Q2" "1990 Q3" "1990 Q4" "1991 Q1" "1991 Q2"
#[7] "1991 Q3" "1991 Q4" "1992 Q1" "1992 Q2" "1992 Q3" "1992 Q4"
NO Addtional packages needed.
If we need a package solution, data.table can be used
library(data.table)
library(stringr)
setDT(df)[, Date:=sub("^(Q.*)", paste0(word(Date[1],1), " \\1") , Date),
cumsum(grepl("^\\d+" , Date))]
df
# Date A B
# 1: 1990 Q1 2 3
# 2: 1990 Q2 4 2
# 3: 1990 Q3 7 6
# 4: 1990 Q4 5 3
# 5: 1991 Q1 7 6
# 6: 1991 Q2 1 8
# 7: 1991 Q3 7 6
# 8: 1991 Q4 9 2
# 9: 1992 Q1 1 7
#10: 1992 Q2 4 6
#11: 1992 Q3 1 3
#12: 1992 Q4 5 8
data
df <- structure(list(Date = c("1990 Q1", "Q2", "Q3", "Q4", "1991 Q1",
"Q2", "Q3", "Q4", "1992 Q1", "Q2", "Q3", "Q4"), A = c(2L, 4L,
7L, 5L, 7L, 1L, 7L, 9L, 1L, 4L, 1L, 5L), B = c(3L, 2L, 6L, 3L,
6L, 8L, 6L, 2L, 7L, 6L, 3L, 8L)), .Names = c("Date", "A", "B"
), row.names = c(NA, -12L), class = "data.frame")

Answer 3

Here is a straight forward way to create the sequence which you are looking for:
numrows<-10 #number of elements desired
#create the sequence of Date objects
qtrseq<-seq(as.Date("1990-01-01"), by="quarter", length.out = numrows)
#created vector for the formatted display
qtrformatted<-paste(as.POSIXlt(qtrseq)$year+1900, quarters(qtrseq))
The downside of this method and the other listed solutions is the lost of the Date object. There is no good way in base R to format the Q1, Q2... and have the object remain a Date object. Depending on your application it might be best to store the date sequence in the data frame and use the statement for qtr formatted only output purposes.
Best of luck.

Answer 4

Assuming Date is a single character column, here's an option using tidyr:
library(tidyr)
# separate date into year and quarter, inserting NAs in year as necessary
df %>% separate(Date, into = c('year', 'quarter'), fill = 'left') %>%
# fill NAs with previous value
fill(year) %>%
# join year and quarter back into a single column
unite(Date, year, quarter, sep = ' ')
# Date A B
# 1 1990 Q1 2 3
# 2 1990 Q2 4 2
# 3 1990 Q3 7 6
# 4 1990 Q4 5 3
# 5 1991 Q1 7 6
# 6 1991 Q2 1 8
# 7 1991 Q3 7 6
# 8 1991 Q4 9 2
# 9 1992 Q1 1 7
# 10 1992 Q2 4 6
# 11 1992 Q3 1 3
# 12 1992 Q4 5 8
Data
df <- structure(list(Date = structure(c(1L, 4L, 5L, 6L, 2L, 4L, 5L,
6L, 3L, 4L, 5L, 6L), .Label = c("1990 Q1", "1991 Q1", "1992 Q1",
"Q2", "Q3", "Q4"), class = "factor"), A = c(2L, 4L, 7L, 5L, 7L,
1L, 7L, 9L, 1L, 4L, 1L, 5L), B = c(3L, 2L, 6L, 3L, 6L, 8L, 6L,
2L, 7L, 6L, 3L, 8L)), .Names = c("Date", "A", "B"), class = "data.frame", row.names = c(NA,
-12L))

Answer 5

Here is something you can try
library(dplyr); library(stringr); library(zoo)
df %>% mutate(Date = paste(na.locf(str_extract(Date, "^[0-9]{4}")),
str_extract(Date, "Q[1-4]$"), sep = " "))
Date A B
1 1990 Q1 2 3
2 1990 Q2 4 2
3 1990 Q3 7 6
4 1990 Q4 5 3
5 1991 Q1 7 6
6 1991 Q2 1 8
7 1991 Q3 7 6
8 1991 Q4 9 2
9 1992 Q1 1 7
10 1992 Q2 4 6
11 1992 Q3 1 3
12 1992 Q4 5 8