brackets: Extract information from a multi-value SOP/DNF expression

In admisc: Adrian Dusa's Miscellaneous

Extract information from a multi-value SOP/DNF expression

Description

Functions to extract information from an expression written in SOP - sum of products form, (or from the canonical DNF - disjunctive normal form) for multi-value causal conditions. It extracts either the values within brackets, or the causal conditions' names outside the brackets.

Usage

betweenBrackets(x, type = "[", invert = FALSE, regexp = NULL)
outsideBrackets(x, type = "[", regexp = NULL)
curlyBrackets(x, outside = FALSE, regexp = NULL)
squareBrackets(x, outside = FALSE, regexp = NULL)
roundBrackets(x, outside = FALSE, regexp = NULL)

Arguments

x	A DNF/SOP expression.
type	Brackets type: curly, round or square.
invert	Logical, if activated returns whatever is not within the brackets.
outside	Logical, if activated returns the conditions' names outside the brackets.
regexp	Optional regular expression to extract information with.

Details

Expressions written in SOP - sum of products are used in Boolean logic, signaling a disjunction of conjunctions.

These expressions are useful in Qualitative Comparative Analysis, a social science methodology that is employed in the context of searching for causal configurations that are associated with a certain outcome.

They are also used to draw Venn diagrams with the package venn, which draws any kind of set intersection (conjunction) based on a custom SOP expression.

The functions curlyBrackets, squareBrackets and roundBrackets are just special cases of the functions betweenBrackets and outsideBrackets, using the argument type as either "{", "[" or "(".

The function outsideBrackets itself can be considered a special case of the function betweenBrackets, when it uses the argument invert = TRUE.

SOP expressions are usually written using curly brackets for multi-value conditions but to allow the evaluation of unquoted expressions, they first needs to get past R's internal parsing system. For this reason, multi-value conditions in unquoted expresions should use the square brackets notation, and conjunctions should always use the product * sign.

Sufficiency is recognized as "=>" in quoted expressions but this does not pass over R's parsing system in unquoted expressions. To overcome this problem, it is best to use the single arrow "->" notation. Necessity is recognized as either "<=" or "<-", both being valid in quoted and unquoted expressions.

Author(s)

Adrian Dusa

Examples

sop <- "A[1] + B[2]*C[0]"

betweenBrackets(sop) # 1, 2, 0

betweenBrackets(sop, invert = TRUE) # A, B, C

# unquoted (valid) SOP expressions are allowed, same result
betweenBrackets(A[1] + B[2]*C[0]) # the default type is "["

# curly brackets are also valid in quoted expressions
betweenBrackets("A{1} + B{2}*C{0}", type = "{")

# or
curlyBrackets("A{1} + B{2}*C{0}")

# and the condition names
curlyBrackets("A{1} + B{2}*C{0}", outside = TRUE)

squareBrackets(A[1] + B[2]*C[0]) # 1, 2, 0

squareBrackets(A[1] + B[2]*C[0], outside = TRUE) # A, B, C

admisc documentation built on Sept. 12, 2024, 6:27 a.m.