R/editEMLfunctions.R
set_missing_data.RdMissing data must have a missing data code and missing data code definition. set_missing_data() can add a single missing value code and single missing value code definition. Missing data should be clearly indicated in the data with a missing data code (e.g "NA", "NaN", "Missing", "blank" etc.). It is generally a good idea to not use special characters for missing data codes (e.g. N/A is not advised). If it is absolutely necessary to leave a cell empty with no code, that cell still needs a missing value code and definition in the metadata. Acceptable codes in this case are "empty" and "blank" with a suitable definition that states the cells are purposefully left empty.
set_missing_data(
eml_object,
files,
columns,
codes,
definitions,
force = FALSE,
NPS = TRUE
)is an EML-formatted R object, either generated in R or imported (typically from an EML-formatted .xml file) using EML::read_eml(
String or List of strings. These are the files that contain undocumented missing data, e.g "my_data_file1.csv".
String or List of strings. These are the columns with missing data for which you would like to add missing data codes and explanations in to the metadata, e.g. "scientificName".
String or list of strings. These are the missing value codes you would like associated with the column in question, e.g. "missing" or "NA".
String or list of strings. These are the missing value code definitions associated with the missing value codes, e.g "not recorded" or "sample damaged when the lab flooded".
logical. Defaults to false. If set to FALSE, a more interactive version of the function requesting user input and feedback. Setting force = TRUE facilitates scripting.
Logical. Defaults to TRUE. Most NPS users should leave this as the default. Only under specific circumstances should it be set to FALSE: if you are not publishing with NPS, if you need to set the publisher location to some place other than the Fort Collins Office (e.g. you are NOT working on a data package) or your product is "for" the NPS but not "by" the NPS and you need to specify a different agency, set NPS = FALSE. When NPS=TRUE, the function will over-write existing publisher info and inject NPS as the publisher along the the Central Office in Fort Collins as the location. Additionally, it sets the "for or by NPS" field to TRUE and specifies the originating agency as NPS.
eml_object
The set_missing_data() be used on an individual column or can accept lists of files, column names, codes, and definitions. Make sure that each missing value has a file, column, single code, and single definition associated with it (if you need multiple missing value codes and definitions per column, please use the set_more_missing() function). If you have many missing value codes and definitions, you might consider constructing (or import) a dataframe to describe them:
Example data frame: df <- data.frame(files = c("table1.csv", "table2.csv", "table3.csv", "table4.csv"), columns = c("EventDate", "scientificName", "eventID", "decimalLatitude"), codes = c("NA", "NA", "missing", "blank"), definitions = c("not recorded", "not identified", "not recorded", "intentionally left blank - not recorded"))
meta2 <- set_missing_data(eml_object = metadata, files = df$files, columns = df$columns, codes = df$codes, definitions = df$definitions)
if (FALSE) { # \dontrun{
#For a single column of data in a single file:
meta2 <- set_missing_data(my_metadata,
"table1.csv",
"scientificName",
"NA",
"Unable to identify")
#For multiple columns of data, potentially across multiple files:
#(blank cells must have the missing value code of "blank" or "empty")
meta2 <- set_missing_data(my_metadata,
files = c("table1.csv", "table1.csv", "table2.csv"),
columns = c("date", "time", "scientificName"),
codes = c("NA", "missing", "blank"),
definitions = c("date not recorded",
"time not recorded",
"intentionally left blank - missing")
)
} # }