convert2regex: Create "REGEXP_LIKE" SQL commands
In songxh0424/MPrOVE-package: Helper functions for appropriateness metrics

Description Usage Arguments Details Value See Also Examples

This function creates a series of "REGEXP_LIKE" SQL commands, given a set of starting strings of ICD codes and a starting pattern to match.

1	convert2regex(nm, str, colName, add.periods = FALSE)

`nm`	A character vector of starting strings of ICD10 codes.
`str`	A specific starting string that is used to match the ones in `nm`. Only the matching ones in `nm` are considered to create the output SQL commands.
`colName`	The corresponding column name to put in "REGEXP_LIKE" commands.
`add.periods`	A logical value controlling whether or not to add periods after the third character of ICD codes.

str is usually a very short string. The function first extracts all the strings in nm that start with str. Then those strings are divided into groups that share the first 1, 2, or 3 letters, with only the last letter being different. For each of these groups, a SQL command like this is created, "OR REGEXP_LIKE (DX_CD,'^I0[1,2,5,6,7,8]','im')".

A character string of a series of "REGEXP_LIKE" SQL commands.

writeSQL

bpExclICD9 <- c(1400:2089,2300:2399,8000:8399,8500:8549,8600:8699,
                9050:9099,92611,92612,929,9520:9529,9580:9599,3040:3042,
                3044,3054:3057,34460,7292,4210,4211,4219,0380:0389,01,
                7300:7309,7832,78079,7808,2859)
bpExclICD10 <- getICD10(bpExclICD9)
codeList <- convertToLike(codes = bpExclICD10, dict = ICD9to10$ICD10,2,4)
nm <- names(codeList$LIKE)

cat(convert2regex(nm,'C',colName = 'DX_CD'))