Basic REDCapR Operations

This vignette covers the the basic functions exposed by the httr and curl packages which allow you to interact with REDCap through its API.

Reading REDCap Data

The function redcap_read_oneshot uses the httr package to call the REDCap API.

report_render_start_time <- Sys.time()

library(knitr)
library(magrittr)
requireNamespace("kableExtra")

opts_chunk$set(
  comment = NA, 
  tidy    = FALSE
)

knit_print.data.frame = function(x, ...) {
  # Adapted from https://cran.r-project.org/web/packages/knitr/vignettes/knit_print.html
  # res = paste(c("", "", kable(x)), collapse = "\n")
  # asis_output(res)
  x %>% 
    # dplyr::mutate_if(
    #   is.character,
    #   function( s ) gsub("\\n", "<br/>", s)
    # ) %>%
    kable(
      col.names = gsub("_", " ", colnames(.)),
      format = "html"
    ) %>%
    kableExtra::kable_styling(
      bootstrap_options = c("striped", "hover", "condensed", "responsive"),
      full_width        = FALSE
    ) %>%
    c("", "", .) %>% 
    paste(collapse = "\n") %>% 
    asis_output()

}
# options(markdown.HTML.header = system.file("misc", "vignette.css", package = "knitr"))
# options(markdown.HTML.header = system.file("misc", "vignette.css", package = "REDCapR"))
# options(markdown.HTML.header = file.path(devtools::inst("REDCapR"), "misc", "vignette.css"))

# options(width=120) #So the output is 50% wider than the default.

Set project-wide values.

There is some information that is specific to the REDCap project, as opposed to an individual operation. This includes the (1) uri of the server, and the (2) token for the user's project.

library(REDCapR) #Load the package into the current R session.
uri <- "https://bbmc.ouhsc.edu/redcap/api/"
token <- "9A81268476645C4E5F03428B8AC3AA7B" #`UnitTestPhiFree` user and simple project (pid 153)

Read all records and fields.

If no information is passed about the desired records or fields, then the entire data set is returned. Only two parameters are required, redcap_uri and token. Unless the verbose parameter is set to FALSE, a message will be printed on the R console with the number of records and fields returned.

#Return all records and all variables.
ds_all_rows_all_fields <- redcap_read(redcap_uri=uri, token=token)$data
ds_all_rows_all_fields #Inspect the returned dataset

Read a subset of the records.

If only a subset of the records is desired, the two approaches are shown below. The first is to pass an array (where each element is an ID) to the records parameter. The second is to pass a single string (where the elements are separated by commas) to the records_collapsed parameter.

The first format is more natural for more R users. The second format is what is expected by the REDCap API. If a value for records is specified, but records_collapsed is not specified, then redcap_read_oneshot automatically converts the array into the format needed by the API.

#Return only records with IDs of 1 and 3
desired_records_v1 <- c(1, 3)
ds_some_rows_v1 <- redcap_read(
   redcap_uri = uri, 
   token      = token, 
   records    = desired_records_v1
)$data

#Return only records with IDs of 1 and 3 (alternate way)
desired_records_v2 <- "1, 3"
ds_some_rows_v2 <- redcap_read(
   redcap_uri        = uri, 
   token             = token, 
   records_collapsed = desired_records_v2
)$data

ds_some_rows_v2 #Inspect the returned dataset

Read a subset of the fields.

If only a subset of the fields is desired, then two approaches exist. The first is to pass an array (where each element is an field) to the fields parameter. The second is to pass a single string (where the elements are separated by commas) to the fields_collapsed parameter. Like with records and records_collapsed described above, this function converts the more natural format (ie, fields) to the format required by the API (ie, fields_collapsed) if fields is specified and fields_collapsed is not.

#Return only the fields record_id, name_first, and age
desired_fields_v1 <- c("record_id", "name_first", "age")
ds_some_fields_v1 <- redcap_read(
   redcap_uri = uri, 
   token      = token, 
   fields     = desired_fields_v1
)$data

#Return only the fields record_id, name_first, and age (alternate way)
desired_fields_v2 <- "record_id, name_first, age"
ds_some_fields_v2 <- redcap_read(
   redcap_uri       = uri, 
   token            = token, 
   fields_collapsed = desired_fields_v2
)$data

ds_some_fields_v2 #Inspect the returned dataset

Read a subset of records, conditioned on the values in some variables.

The two techniques above can be combined when your datasets are large and you don't want to pull records with certain values. Suppose you want to select subjects from the previous dataset if the were born before 1960 and their weight was over 70kg. Two calls to the server are required. The first call to REDCap pulls all the records, but for only three columns: record_id, dob, and weight. From this subset, identify the records that you want to pull all the data for; in this case, the desired record_id values are 3 & 5. The second call to REDCap pulls all the columns, but only for the identified records.

######
## Step 1: First call to REDCap
desired_fields_v3 <- c("record_id", "dob", "weight")
ds_some_fields_v3 <- redcap_read(
   redcap_uri = uri, 
   token      = token, 
   fields     = desired_fields_v3
)$data

ds_some_fields_v3 #Examine the these three variables.

######
## Step 2: identify desired records, based on age & weight
before_1960 <- (ds_some_fields_v3$dob <= as.Date("1960-01-01"))
heavier_than_70_kg <- (ds_some_fields_v3$weight > 70)
desired_records_v3 <- ds_some_fields_v3[before_1960 & heavier_than_70_kg, ]$record_id

desired_records_v3 #Peek at IDs of the identified records

######
## Step 3: second call to REDCap
#Return only records that met the age & weight criteria.
ds_some_rows_v3 <- redcap_read(
   redcap_uri = uri, 
   token      = token, 
   records    = desired_records_v3
)$data

ds_some_rows_v3 #Examine the results.

Additional Returned Information

The examples above have shown only the resulting data.frame, by specifying $data at the end of the call. However, more is available to those wanting additional information, such as:

  1. The data object has the data.frame, as in the previous examples.
  2. The success boolean value indicates if redcap_read_oneshot believes the operation completed as intended.
  3. The status_codes is a collection of http status codes, separated by semicolons. There is one code for each batch attempted.
  4. The outcome_messages: A collection of human readable strings indicating the operations' semicolons. There is one code for each batch attempted. In an unsuccessful operation, it should contain diagnostic information.
  5. The records_collapsed field passed to the API. This shows which record subsets, if any, were requested.
  6. The fields_collapsed fields passed to the API. This shows which field subsets, if any, were requested.
  7. The elapsed_seconds measures the duration of the call.
#Return only the fields record_id, name_first, and age
all_information <- redcap_read(
   redcap_uri = uri, 
   token      = token, 
   fields     = desired_fields_v1
)
all_information #Inspect the additional information

Session Information

For the sake of documentation and reproducibility, the current report was rendered in the following environment. Click the line below to expand.

Environment

if( requireNamespace("devtools", quietly = TRUE) ) {
  devtools::session_info()
} else {
  sessionInfo()
} 

report_render_duration_in_seconds <- round(as.numeric(difftime(Sys.time(), report_render_start_time, units="secs")))

Report rendered by r Sys.info()["user"] at r strftime(Sys.time(), "%Y-%m-%d, %H:%M %z") in r report_render_duration_in_seconds seconds.



Try the REDCapR package in your browser

Any scripts or data that you put into this service are public.

REDCapR documentation built on May 29, 2017, 9:24 a.m.