Extract portion of taxonomic string (i.e., kingdom, phyla, etc.)

[lsandell]

Hi there! I have taxonomic data with the phyla through genus listed from microbiome 16s sequencing. I want to pull just the 'Family' name out of the string, but each character string is different due to the nature of each taxonomy identified within the sample. 

I have included a screenshot of the character variable "taxon" that has this string information. It's a character variable with 106 length. (I cannot share the data in full so I hope the image is enough context). Is there an approach you'd recommend taking to pull out this substring from variable starting positions? I do have a delimiter of sorts - the " / ". The most efficient method would be useful as I have over 1 million records.

Thank you!

[sbxkoenk]

---

@lsandell wrote:

I want to pull just the 'Family' name out of the string, ...

---

What is the 'Family' name in the strings listed?

What is the substring you need as an outcome after parsing every taxon string?

Koen

[lsandell]

Just one word from the whole string (what's in between the slashes). The target string is the name before the final forward slash (/) [i.e., second to last alphabetic portion].

[SASKiwi]

So what is your business rule for identifying the family name in the string?

[lsandell]

Thanks! It's the name before the final forward slash (/) [i.e., second to last alphabetic portion].

[SASKiwi · Accepted solution]

Try this. The SCAN function is finding the second word reading from the right:

data test;
  string = 'w1/w2/w3/w4/w5/w6';
  Family = scan(string,-2,'/');
  put _all_;
run;