I am doing big data analysis, and I want to add up each individual's physical activity. pha_04z1 is the number of days of vigorous physical activity in the last week, and pha_05z1 and pha_06z1 are hours and minutes. pha_07z1 is the number of days of moderate physical activity in a week, and pha_08z1 and pha_09z1 are hours and minutes. Although it is not in raw-data, I am trying to create variables called ph_a0100, ph_0200, ph_0300, ph_0400, ph_0500 to obtain the final physical activity amount. Moderate or more physical activity is defined as 'in the past week, 20 minutes or more per day and 3 days a week or more' is defined. The SAS codes for this are as follows.

if 0<=pha_05z1 <=24 and pha_06z1=. then do;
 ph_a0100=pha_05z1*60;
end;
if a<=pha_05z1<=24 and 0<=pha_06z1 <=59 then do;
 ph_a0100=(pha_05z1*60) pha_06z1;
end;
if pha_05z1=. and 0<=pha_06z1<=59 then do;
 ph_a0100=pha_06z1;
end;
if pha_04z1 in (0:2) the ph_a0200=0;
else if pha_04z1 in (3:7) then do;
 if ph_a0100=. then ph_a0200=.;
 else if 0<=ph_a0100<=19 then ph_a0200=0;
 else if 20<=ph_a0100 then ph_a0200=1;
end;
-----
(Abstinence from vigorous physical activity)
-----
if ph_a0200=1 or ph_a0400=1 then ph_a0500=1;
else if ph_a0200=0 and ph_a0400=0 then ph_a0500=0;

Below is what I made in R code.

if (pha_05z1<=24 & pha_05z1>=0 & pha_06z1==88)
{
  ch2020_$ph_a0100 <- pha_05z1*60
}
if (pha_05z1<=24 & pha_05z1>=0 & pha_06z1<=59 & pha_06z1>=0)
{
  ch2020_$ph_a0100 <- pha_05z1*60   pha_06z1
}
if (pha_05z1==88 & pha_06z1<=59 & pha_06z1>=0)
{
  ch2020_$ph_a0100 <- pha_06z1
}
ch2020_$ph_a0200 <- 
  ifelse(pha_04z1%in%c(0,1,2),0, 
         ifelse(pha_04z1>=3 & ch2020_$ph_a0100==NA),NA, 
         ifelse(ch2020_$ph_a0100<=19 & ch2020_$ph_a0100 >=0),0,1)

This code doesn't work. How can I solve this? Thank you in advance. Please help.

CodePudding user response：

Lacking data, I'll throw out some code, perhaps it'll work.

1. Realize that while in SAS, you can do a conditional reassignment with your if statements, in R the <- operator overwrites everything unless the LHS is an indexed reference (using [ or $). (I do not know SAS well, so I could be mistaken in the interpretation of your code.)

2. Don't use single-& in an if statement unless it is wrapped in logical aggregators such as any, all, etc. While it currently works, the premise of R's if operator is that its condition must be length exactly 1, anything else is a mistake (and, in R-4.2.0, will lead to an error, not just a warning).

3. I'm inferring that your references to pha_06z1 and the like are actually columns within ch2020_. Lack of data makes this hard to know for sure.

Three ways:

1. Nested ifelse:

    ch2020_$ph_a0100 <- ifelse(ch2020_$pha_05z1<=24 & ch2020_$pha_05z1>=0 & ch2020_$pha_06z1==88, 
                               ch2020_$pha_05z1*60,
                               ifelse(ch2020_$pha_05z1<=24 & ch2020_$pha_05z1>=0 & ch2020_$pha_06z1<=59 & ch2020_$pha_06z1>=0,
                                      ch2020_$pha_05z1*60   ch2020_$pha_06z1,
                                      ifelse(ch2020_$pha_05z1==88 & ch2020_$pha_06z1<=59 & ch2020_$pha_06z1>=0,
                                             ch2020_$pha_06z1, ch2020_$ph_a0100)))
   

2. Nested ifelse, but wrapping it in with to make it a little more readable:

   ch2020_$ph_a0100 <- with(ch2020_,
     ifelse(pha_05z1<=24 & pha_05z1>=0 & pha_06z1==88, 
            pha_05z1*60,
            ifelse(pha_05z1<=24 & pha_05z1>=0 & pha_06z1<=59 & pha_06z1>=0,
                   pha_05z1*60   pha_06z1,
                   ifelse(pha_05z1==88 & pha_06z1<=59 & pha_06z1>=0,
                          pha_06z1, ph_a0100)))
   )
   

3. Assign a default value, then iteratively replace sub-indexed portions.

   # ch2020_$ph_a0100 is predefined with some value or just NA
   ind <- with(ch2020_, pha_05z1<=24 & pha_05z1>=0 & pha_06z1==88)
   ch2020_$ph_a0100[ind] <- ch2020_$pha_05z1[ind] * 60
   ind <- with(ch2020_, pha_05z1<=24 & pha_05z1>=0 & pha_06z1<=59 & pha_06z1>=0)
   ch2020_$ph_a0100[ind] <- with(ch2020_, pha_05z1[ind]*60   pha_06z1[ind])
   ind <- with(ch2020_, pha_05z1==88 & pha_06z1<=59 & pha_06z1>=0)
   ch2020_$ph_a0100[ind] <- ch2020_$pha_06z1