extract specific substring sql

Question

I have a database that contains 16 columns. In the 16 column I have the following text:

ASN_MAF=0.09;DOMAINS=Pfam_domain:PF00168,Prints_domain:.

I want to extract PF00168, so the substring between Pfam_domain: and ,. All rows have this pattern: Pfam_domain: and ,.

I try to do this request but it doesn't work:

res = sqldf(" 
SELECT SUBSTRING(v16, CHARINDEX("Pfam_domain:",v16)+1, 10000), CHARINDEX(",",v16)-1 )
FROM GeminiTable_germ
        ")

Answer 1

Try:

SELECT SUBSTRING(v16, CHARINDEX('Pfam_domain:', v16)+12, CHARINDEX         
(',',v16) - (CHARINDEX('Pfam_domain:', v16)+12))

Note that I changed the " to ' inside the SQL statement - that is an important distinction in SQL.

Answer 2

Assuming we want the string between colon and comma, instr(v16, ':')+1 is the character position after the colon. Also, the string we want is of length instr(v16, ',') - instr(v16, ':')-1 so use substr with those 2nd and 3rd arguments:

library(sqldf)

GeminiTable_germ <- data.frame(v16 =
  "ASN_MAF=0.09;DOMAINS=Pfam_domain:PF00168,Prints_domain:.")

sqldf("select substr(v16, instr(v16, ':')+1, instr(v16, ',')-instr(v16, ':')-1) v16new 
  from GeminiTable_germ")

giving:

   v16new
1 PF00168

We could alternately break it up and write it like this instead:

field <- function(x, from, to) {
      from_pos <- sprintf("instr(%s, '%s')+1", x, from)
      to_pos <- sprintf("instr(%s, '%s')-%s-2", x, to, from_pos)
      sprintf("substr(%s, %s, %s)", x, from_pos, to_pos)
}
field('v16', ':', ',') # view generated code

fn$sqldf("select `field('v16', ':', ',')` v16new from GeminiTable_germ")