Error fetching resolutiondate

[matbmeijer]

Error fetching resolutiondate

[matbmeijer]

Hi @hannessnogren, my apologies for the late response - I come back to this always when I have time. I will check your issue and come back to you.

[Dschaykib]

I had the same issue for the duedate.
After looking a bit into your function I found the problem:

Let's say you have two issues and one does not have an duedate (or resolutiondate or any other field). Within JiraJSON2df you check for the format of the field and set is to NA if it is missing. In the next step you use data.table::rbindlist() and here is the error.

The two data.tables are combined, but one has a date column and the other one a logical.

Maybe a solution would be to check all fields and set them to the most used class?

[Dschaykib]

UPDATE:
I added some data.table::as.data.table and fiexd some issues.

Here is an approach to set columns, that have different classes all to characters:

DT1 <- data.table(VAR1 = "a", VAR2 = as.Date("2020-04-02"))
DT2 <- data.table(VAR1 = "b", VAR2 = NA)
DT3 <- data.table(VAR1 = "c", VAR2 = as.Date("2020-04-02"))

check_dt <- function(input = list()) {
  classes <- lapply(input, function(dt) {
    dt <- data.table::as.data.table(dt)
    as.data.table(t(dt[, sapply(.SD, class), .SDcol = names(dt)]))}
    )
  
  classes_dt <- rbindlist(classes)
  
  idx <- classes_dt[, sapply(.SD, uniqueN), .SDcol = names(classes_dt)] > 1
  
  idx_cols <- names(idx)[idx]
  
  output <- lapply(input, function(dt, idx_cols){
    dt <- data.table::as.data.table(dt)
    these_cols <- intersect(names(dt), idx_cols)
    if (length(these_cols) >0 ) {
      dt[, c(these_cols) := lapply(.SD, as.character), .SDcol = these_cols]
    }
    return(dt)
    },
    idx_cols = idx_cols)
  return(output)
}

rbindlist(check_dt(input = list(DT1, DT2, DT3)))

> out
   VAR1       VAR2
1:    a 2020-04-02
2:    b       <NA>
3:    c 2020-04-02

[matbmeijer]

Hi @hannessnogren & @Dschaykib, I changed substantially the internals of the package to increase the robustness of the package and to give a better support for JQL fields. With the latest commit 3452c86 Now your queries should work - here an example with the field "resolutiondate":

JiraQuery2R(domain = "https://bitvoodoo.atlassian.net",
            query = "project='CONGRATS'",
            fields = c("resolutiondate")) %>%
  head() %>% knitr::kable()

id	self	key	ID	resolutiondate
57849	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57849	CONGRATS-39	1	NA
57848	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57848	CONGRATS-38	2	NA
57847	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57847	CONGRATS-37	3	NA
57810	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57810	CONGRATS-36	4	2019-10-29 09:32:08
57667	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57667	CONGRATS-32	5	2019-09-23 16:33:36
57502	https://bitvoodoo.atlassian.net/rest/api/latest/issue/57502	CONGRATS-30	6	2019-05-20 09:00:11

@Dschaykib - I tried your commit, and while it solved some of the issues (it did not work always), I preferred to rethink the internal logic instead of repatching the issue. But thank you very much for your help, it helped me quite a lot.