With this code, I can only get ASRIs for 2021, how can I get all the ASRIs from 1990 to 2021 at once?

[yuyu11]

EC_2021_ASPR <- subset(EC,EC$year==2021 & 
                         EC$location=="Global"&
                           EC$age_name=='Age-standardized' & 
                         EC$sex_name=="Both"&
                           EC$metric_name== 'Rate' &
                           EC$measure_name=='Incidence')
EC_2021_ASPR$val <- round(EC_2021_ASPR$val,1)
EC_2021_ASPR$lower <- round(EC_2021_ASPR$lower,1)
EC_2021_ASPR$upper <- round(EC_2021_ASPR$upper,1)
EC_2021_ASPR$lower_to_upper <- paste(EC_2021_ASPR$lower,EC_2021_ASPR$upper,sep = ' to ')
EC_2021_ASPR$lower_to_upper <- paste('(',EC_2021_ASPR$lower_to_upper,')',sep = '')
EC_2021_ASPR$ASPR <- paste(EC_2021_ASPR$val,EC_2021_ASPR$lower_to_upper,sep = ' ')
print(EC_2021_ASPR)

[lorae]

You haven’t provided a data sample, so I can only base my answer on the code and assumptions about your dataset. (If you provide a data sample, we can test the code). It seems like you're working with a data frame (or tibble) called EC and creating EC_2021_ASPR as a filtered version.

The simplest change is just to stop selecting for year==2021 in the first line. Here, I've modified the code to say EC$year >= 1990 & EC$year <= 2021. If the years 1990 through 2021 are the only years available in your sample, then this entire line may not be necessary at all - you could simply remove it.

ASPR_all_years <- subset(EC, 
                         EC$year >= 1990 & EC$year <= 2021 &
                         EC$location == "Global" &
                         EC$age_name == "Age-standardized" &
                         EC$sex_name == "Both" &
                         EC$metric_name == "Rate" &
                         EC$measure_name == "Incidence")

But your code uses base R (not the newest conventions) and is maybe a bit messier than it could be - might I suggest a few changes?

Firstly, you can use the dplyr package, which is very commonly used in R nowadays, instead of the base R functions you're using. This allows you to filter the data along all the same criteria without having to repeat the EC$ expression numerous times:

# Un-comment this line and run once if you don't already have dplyr installed
# install.packages("dplyr")
library(dplyr)

ASPR_all_years <- EC |>
  filter(year %in% 1990:2021,
         location == "Global",
         age_name == "Age-standardized",
         sex_name == "Both",
         metric_name == "Rate",
         measure_name == "Incidence")

The pipe symbol (|>) lets you string together commands. Basically, this code says "Define ASPR_all_years as EC. Now take only the rows where the year falls in a range from 1990 to 2021 (inclusive), location is 'Global', ... etc."

The other advantage to using the dplyr package is that you can avoid continuously redefining the same data frame. In the following lines of code, you continuously make incremental tweaks to EC_2021_ASPR. This can get clunky. Instead, I'd suggest using the mutate function, which is also from the dplyr package:

ASPR_all_years <- EC |>
  # Filter along needed criteria. Can remove the first line if 1990 - 2021 are the only years in the data
  filter(
    year %in% 1990:2021,
    location == "Global",
    age_name == "Age-standardized",
    sex_name == "Both",
    metric_name == "Rate",
    measure_name == "Incidence"
    ) |>
  # Use the mutate function to round val, lower, and upper and to create lower_to_upper string
  mutate(
    val = round(val, 1),
    lower = round(lower, 1),
    upper = round(upper, 1),
    lower_to_upper = paste0(val, " (", lower, " to ", upper, ")")
  )

Here we're saying "Define ASPR_all_years as EC. Now take only the rows where the year falls in a range from 1990 to 2021 (inclusive), location is 'Global', ... etc. Now redefine the val, lower, and upper columns to be rounded to the tenths place and create a lower_to_upper column which is a specially formatted string with the number followed by its range in parentheses."

Good luck!