Thelibname() Function

Thelibrlibname() function is quitesimilar to the SAS®libname statement. The first parameteris the name of the library. The second parameter is a path to adirectory the library will point to. The third parameter is the enginewith which to read and write the data.

library(libr)# Get path to sample datapkg<-system.file("extdata",package ="libr")# Define data librarylibname(sdtm, pkg,"csv")

Thelibname() function above will send two types ofinformation to the console:

• The column specifications for each data file imported
• A summary print-out of the library

The summary print-out looks like this:

# library 'sdtm': 8 items- attributes: csv not loaded- path: C:/packages/libr/inst/extdata- items:  Name Extension Rows Cols     Size        LastModified1   AE       csv  150   27  88.1 Kb 2020-09-18 14:30:232   DA       csv 3587   18 527.8 Kb 2020-09-18 14:30:233   DM       csv   87   24  45.1 Kb 2020-09-18 14:30:234   DS       csv  174    9  33.7 Kb 2020-09-18 14:30:235   EX       csv   84   11    26 Kb 2020-09-18 14:30:236   IE       csv    2   14    13 Kb 2020-09-18 14:30:237   SV       csv  685   10  69.9 Kb 2020-09-18 14:30:248   VS       csv 3358   17   467 Kb 2020-09-18 14:30:24

The summary displays what type of library it is, where it is located,and what data (if any) is already in the library directory. In thiscase, there are eight ‘csv’ files available.

For each of the eight files, thelibname() function alsodisplayed the column specifications used to import the data file. Acolumn specification looks like this:

$VS-- Column specification ------------------------------------------cols(  STUDYID = col_character(),  DOMAIN = col_character(),  USUBJID = col_character(),  VSSEQ = col_double(),  VSTESTCD = col_character(),  VSTEST = col_character(),  VSPOS = col_character(),  VSORRES = col_double(),  VSORRESU = col_character(),  VSSTRESC = col_double(),  VSSTRESN = col_double(),  VSSTRESU = col_character(),  VSBLFL = col_character(),  VISITNUM = col_double(),  VISIT = col_character(),  VSDTC = col_date(format = ""),  VSDY = col_double())

The column specification shows how the data was imported. Since ‘csv’files do not contain well-defined data type information on each of thecolumns, thelibname() function has to guess at the datatypes. The column specification shows you what the guesses were. This isuseful information. You should review these column specifications to seeif thelibname() function guessed correctly. If it did notguess correctly, you can control the import data types by sending aspecs() collection ofimport_spec() objects totheimport_specs parameter on thelibname()function. See thespecs() documentation for an example andadditional details.

Accessing Data

Observe that there is difference between the SAS®libname statement and thelibrlibname() function. The difference is that after the SAS®libname statement is called, the data is immediatelyavailable to your code using two-level (<library>.<dataset>)syntax.

With thelibr function, on the other hand, the datais immediately available using list syntax on the library variable name.That means you can get to your data using the dollar sign ($), likethis:

# View a datasetsdtm$DM# # A tibble: 87 × 24#    STUDYID DOMAIN USUBJID SUBJID RFSTDTC    RFENDTC RFXSTDTC RFXENDTC RFICDTC    RFPENDTC#    <chr>   <chr>  <chr>   <chr>  <date>     <date>  <lgl>    <lgl>    <date>     <date>#  1 ABC     DM     ABC-01… 049    2006-11-07 NA      NA       NA       2006-10-25 NA#  2 ABC     DM     ABC-01… 050    2006-11-02 NA      NA       NA       2006-10-25 NA#  3 ABC     DM     ABC-01… 051    2006-11-02 NA      NA       NA       2006-10-25 NA#  4 ABC     DM     ABC-01… 052    2006-11-06 NA      NA       NA       2006-10-31 NA#  5 ABC     DM     ABC-01… 053    2006-11-08 NA      NA       NA       2006-11-01 NA#  6 ABC     DM     ABC-01… 054    2006-11-16 NA      NA       NA       2006-11-07 NA#  7 ABC     DM     ABC-01… 055    2006-12-06 NA      NA       NA       2006-10-31 NA#  8 ABC     DM     ABC-01… 056    2006-11-28 NA      NA       NA       2006-11-21 NA#  9 ABC     DM     ABC-01… 113    2006-12-05 NA      NA       NA       2006-11-28 NA# 10 ABC     DM     ABC-01… 114    2006-12-14 NA      NA       NA       2006-12-01 NA# #  77 more rows# #  14 more variables: DTHDTC <lgl>, DTHFL <lgl>, SITEID <chr>, BRTHDTC <date>, AGE <dbl>,# #   AGEU <chr>, SEX <chr>, RACE <chr>, ETHNIC <chr>, ARMCD <chr>, ARM <chr>, ACTARMCD <lgl>,# #   ACTARM <lgl>, COUNTRY <lgl># #  Use `print(n = ...)` to see more rows

Using this syntax, your dataset can be passed into any R function.For example, here we can subset the dataset for a particularsubject:

# Subset the datadat <- subset(sdtm$DM, SUBJID == '050')# View resultsdat# # A tibble: 1 × 24#   STUDYID DOMAIN USUBJID  SUBJID RFSTDTC    RFENDTC RFXSTDTC RFXENDTC RFICDTC    RFPENDTC#   <chr>   <chr>  <chr>    <chr>  <date>     <date>  <lgl>    <lgl>    <date>     <date>  # 1 ABC     DM     ABC-01-… 050    2006-11-02 NA      NA       NA       2006-10-25 NA      # #   14 more variables: DTHDTC <lgl>, DTHFL <lgl>, SITEID <chr>, BRTHDTC <date>, AGE <dbl>,# #   AGEU <chr>, SEX <chr>, RACE <chr>, ETHNIC <chr>, ARMCD <chr>, ARM <chr>, ACTARMCD <lgl>,# #   ACTARM <lgl>, COUNTRY <lgl>

The dollar sign syntax shown above is recommended for the mostmemory-efficient programming. If you are writing production code to berun in batch, use the dollar sign syntax.

Thelib_load() Function

For convenience, the package also provides a way to get two-level dotsyntax, similar to SAS®. To get the dot syntax, you first have to callthelib_load() function.

lib_load(sdtm)# # library 'sdtm': 8 items# - attributes: csv loaded# - path: C:/packages/libr/inst/extdata# - items:#   Name Extension Rows Cols     Size        LastModified# 1   AE       csv  150   27  88.1 Kb 2020-09-18 14:30:23# 2   DA       csv 3587   18 527.8 Kb 2020-09-18 14:30:23# 3   DM       csv   87   24  45.1 Kb 2020-09-18 14:30:23# 4   DS       csv  174    9  33.7 Kb 2020-09-18 14:30:23# 5   EX       csv   84   11    26 Kb 2020-09-18 14:30:23# 6   IE       csv    2   14    13 Kb 2020-09-18 14:30:23# 7   SV       csv  685   10  69.9 Kb 2020-09-18 14:30:24# 8   VS       csv 3358   17   467 Kb 2020-09-18 14:30:24

Notice on the console printout that the library is now “loaded”. Thatmeans the data has been loaded into the workspace, and is availableusing two-level dot syntax. If you are working inRStudio, the environment pane will now show all thedatasets available in the library.

At this point, you can work with your data very much the same way asyou would in SAS®. You can pass these datasets into statisticalfunctions, or manipulate them withdplyr functions.Note that you can also work with individual variables on the datasetsusing dollar sign (“$”) syntax.

# Get total number of recordsnrow(sdtm.DM)# [1] 87# Get frequency counts for each armtable(sdtm.DM$ARM)# ARM A          ARM B          ARM C          ARM D SCREEN FAILURE# 20             21             21             23              2

The datasets will be available in the workspace for the length ofyour session. If you wish to unload them from the workspace, call thelib_unload() function. See thelib_load() andlib_unload() documentation for additional information onthese functions.

To see more examples of thelibr data managementfunctions, refer to the articles onBasicLibrary Operations andLibraryManagement.

Thedictionary() Function

Once you have a library defined, you may want to examine the columnattributes for the datasets in that library. Examining those columnattributes can be accomplished with thedictionary()function. Thedictionary() function returns a tibble ofinformation about the data in the library.

Continuing from the example above, let’s look at the dictionary forthe ‘sdtm’ library created previously.

dictionary(sdtm)# # A tibble: 130 x 10#    Name  Column  Class     Label Description Format Width Justify  Rows   NAs#    <chr> <chr>   <chr>     <chr> <chr>       <lgl>  <int> <chr>   <int> <int>#  1 AE    STUDYID character NA    NA          NA         3 NA        150     0#  2 AE    DOMAIN  character NA    NA          NA         2 NA        150     0#  3 AE    USUBJID character NA    NA          NA        10 NA        150     0#  4 AE    AESEQ   numeric   NA    NA          NA        NA NA        150     0#  5 AE    AETERM  character NA    NA          NA        72 NA        150     0#  6 AE    AELLT   logical   NA    NA          NA        NA NA        150   150#  7 AE    AELLTCD logical   NA    NA          NA        NA NA        150   150#  8 AE    AEDECOD character NA    NA          NA        43 NA        150     0#  9 AE    AEPTCD  numeric   NA    NA          NA        NA NA        150     0# 10 AE    AEHLT   character NA    NA          NA        63 NA        150     0# # ... with 120 more rows

The resulting dictionary table shows the name of the dataset, thecolumn name, and some interesting attributes related to each column. Asyou can see, thelibr dictionary table is overall quitesimilar to a SAS® dictionary table. See thedictionary()function documentation for more information.

age_groups<-datastep(sdtm.DM,keep =c("USUBJID","AGE","AGEG"), {if (AGE>=18& AGE<=29)                           AGEG<-"18 to 29"elseif (AGE>=30& AGE<=44)                           AGEG<-"30 to 44"elseif (AGE>=45& AGE<=59)                           AGEG<-"45 to 59"else                           AGEG<-"60+"                       })age_groups# # A tibble: 87 x 3#    USUBJID      AGE AGEG#    <chr>      <dbl> <chr>#  1 ABC-01-049    39 30 to 44#  2 ABC-01-050    47 45 to 59#  3 ABC-01-051    34 30 to 44#  4 ABC-01-052    45 45 to 59#  5 ABC-01-053    26 18 to 29#  6 ABC-01-054    44 30 to 44#  7 ABC-01-055    47 45 to 59#  8 ABC-01-056    31 30 to 44#  9 ABC-01-113    74 60+# 10 ABC-01-114    72 60+# # ... with 77 more rows