add %tq date format

[JanMarvin]

A possible fix

diff --git a/R/read.R b/R/read.R
index b120e07..8926c73 100644
--- a/R/read.R
+++ b/R/read.R
@@ -341,6 +341,20 @@ read.dta13 <- function(file, convert.factors = TRUE, generate.factors=FALSE,
       z <- z - rowSums(outer(z, ls, ">="))
       as.POSIXct(z, origin = "1960-01-01")
     }
+    
+    convert_dt_q <- function(x) { 
+      z <- x / 4 
+      qrt <- z - floor(z)
+      z <- 1960 + floor(z)
+      
+      z[qrt == 0.00] <- paste0(z[qrt == 0.00], "-1-1")
+      z[qrt == 0.25] <- paste0(z[qrt == 0.25], "-4-1")
+      z[qrt == 0.50] <- paste0(z[qrt == 0.50], "-7-1")
+      z[qrt == 0.75] <- paste0(z[qrt == 0.75], "-10-1")
+      
+      z <- as.Date(z, "%Y-%m-%d")
+      z
+    }
 
     ff <- attr(data, "formats")
     ## dates <- grep("%-*d", ff)
@@ -359,6 +373,7 @@ read.dta13 <- function(file, convert.factors = TRUE, generate.factors=FALSE,
 
     for (v in grep("%tc", ff)) data[[v]] <- convert_dt_c(data[[v]])
     for (v in grep("%tC", ff)) data[[v]] <- convert_dt_C(data[[v]])
+    for (v in grep("%tq", ff)) data[[v]] <- convert_dt_q(data[[v]])
   }
 
   if (convert.factors) {

[JanMarvin]

closed in b442419

[anjoung]

Hi! I am getting the following error message when I use your package:
dta3 <- read.dta13("edited/temp.dta") Warning messages: 1: In convert_dt_q(data[[v]]) : conversion of %tq failed 2: In convert_dt_q(data[[v]]) : conversion of %tq failed

Here's a link to a sample of the data I am working with which replicates the problem. The relevant variable is quarterly_dates I think that the package actually does effectively import quarterly dates in the test case, but since I am producing so many data files I cannot manually check each file every time this warning comes up. Would like to know the root cause and see if there is a fix!

[JanMarvin]

Maybe due to missing values in the date file? Could you have a look @sjewo ? I'm a bit swamped with other things ATM.

[JanMarvin]

Thanks for the report @anjoung , that's not supposed to throw a warning and looks like lazy coding.

[sjewo]

Here's a link to a sample of the data I am working with which replicates the problem.

The file is not public. Could you provide another link?

[anjoung]

Sorry about that. Let me know if this link works: https://www.dropbox.com/s/3s90yn9hmcc6n7z/temp.dta?dl=0

[sjewo]

Thanks for the new link, @anjoung !

The warnings only occurs if NA values are present (e.g. in variable treat_quarter).

I posted a fix in 0a82b55.

Could you try the current testing branch?

library(remotes)
remotes::install_github("sjewo/readstata13", ref="testing")

[JanMarvin]

Thanks!