Coerce session check object to a data frame

S3 as.data.frame() methods for the three classes this package defines, letting each be dropped into ordinary data frame workflows (filtering, joining, export) instead of only being inspected via print()/format().

Usage

# S3 method for class 'sessioncheck_status'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

# S3 method for class 'sessioncheck_sessioncheck'
as.data.frame(x, row.names = NULL, optional = FALSE, ...)

# S3 method for class 'sessioncheck_sessionstate'
as.data.frame(x, row.names = NULL, optional = FALSE, which = "packages", ...)

# S3 method for class 'sessioncheck_sessionstatediff'
as.data.frame(x, row.names = NULL, optional = FALSE, which = "packages", ...)

Arguments

x

An object of class sessioncheck_status, sessioncheck_sessioncheck, sessioncheck_sessionstate, or sessioncheck_sessionstatediff

row.names

Ignored

optional

Ignored

...

Ignored

which

For sessioncheck_sessionstate objects, which tabular component to return: one of "packages" (the default), "globalenv", or "attachments". For sessioncheck_sessionstatediff objects, the same three section names instead select a long-format diff table (see Details); the default is likewise "packages". Ignored for other classes.

Value

A data frame

Details

For sessioncheck_status and sessioncheck_sessioncheck objects, this coercion is lossless: every entity and status recorded in x appears as a row in the result. That guarantee does not extend to sessioncheck_sessionstate objects: sessionstate() captures more than any single rectangular table can hold, mixing scalar fields (platform, locale, matrix, document, machine, git, timing, rng), a bare character vector (libpaths), and three differently-shaped tables (packages, globalenv, attachments). as.data.frame() returns whichever one of those three tables which selects; none of the scalar fields or libpaths are represented in the result. Use x$platform, x$machine, x$git, x$libpaths, etc. (or unclass(x) for everything at once) to access those directly.

sessioncheck_sessionstatediff objects (from compare_sessionstates()) coerce differently again: each which selects a single long-format table with one row per key (package/name/name, for "packages"/"globalenv"/"attachments" respectively) and tracked field, with columns <key>, change ("added", "removed", or "modified"), field, old, and new. A key present in only one snapshot contributes one row per tracked field, all with the same change, and old/new NA on whichever side it didn't exist; a key present in both snapshots contributes a row only for fields that actually changed. The tracked fields are attached/ondisk_version/loaded_version/source for "packages", class/size/hash for "globalenv", and type for "attachments" – the same fields compare_sessionstates() tracks for its own modified tables. "globalenv" additionally has a verified column (NA for "added"/"removed" rows, since there is nothing to verify when a key only exists in one snapshot; TRUE/FALSE for "modified" rows – see compare_sessionstates() for what verified means). "attachments" never has "modified" rows, since a type change for an existing search-path entry isn't a realistic scenario.