Excel: Extract text with FIND and MID
Data analysis begins with usable data, and that means every piece organized nicely into its own field where we can count, sort and otherwise test it out.
What if you get a spreadsheet where the pieces of data are all packed in one field? Say, something like this (which I cobbled together from Major League Baseball data in honor of the Nationals’ first playoff appearance):
NAME: Sean Burnett POS: RP AGE: 30 WT: 200 BORN: Dunedin, FL SALARY: 2350000 NAME: Tyler Clippard POS: RP AGE: 27 WT: 200 BORN: Lexington, KY SALARY: 1650000 NAME: Ross Detwiler POS: SP AGE: 26 WT: 174 BORN: St. Louis, MO SALARY: 485000 NAME: Christian Garcia POS: RP AGE: 27 WT: 215 BORN: Miami, FL SALARY: N/A NAME: Gio Gonzalez POS: SP AGE: 27 WT: 205 BORN: Hialeah, FL SALARY: 3335000 NAME: Mike Gonzalez POS: RP AGE: 34 WT: 215 BORN: Robstown, TX SALARY: N/A NAME: Ryan Mattheus POS: RP AGE: 28 WT: 215 BORN: Sacramento, CA SALARY: 481000 NAME: Craig Stammen POS: RP AGE: 28 WT: 200 BORN: Coldwater, OH SALARY: 485000 NAME: Drew Storen POS: RP AGE: 25 WT: 180 BORN: Indianapolis, IN SALARY: 498750 NAME: Jordan Zimmermann POS: SP AGE: 26 WT: 218 BORN: Auburndale, WI SALARY: 2300000
Let’s say you want to extract the city of each player’s birth into a separate field. The varying length of each player’s name means the birth place isn’t always in the same position in the string, so a typical text-to-columns operation won’t work. So, how to do it?
The answer lies in two very handy Excel functions: FIND and MID.
FIND locates characters you specify and returns its numeric place in the string.
MID returns X characters from a string beginning at a location X you specify.
For example, we can locate the position where each city name begins by using FIND to locate the string “BORN:” in each cell. The city name itself always starts six characters after the position of that string, so we add six to the result:
In the first row above, the functions returns 50. In the second row, 52. We’ll feed that value to the MID function as the starting point for our extraction.
MID takes three arguments: Text or cell, position to start, number of characters to return. So, we use the above FIND function as the second argument and, for now, extract 10 characters:
That gets us part of the way there. We’re starting in the right spot, but 10 characters isn’t always the length of the city and state, so it leads to choppy results:
Dunedin, F Lexington, St. Louis, Miami, FL Hialeah, F Robstown, Sacramento Coldwater, Indianapol Auburndale
What we need to do is tell MID the exact number of characters to return each time even though the length of the city varies. We can figure this out using FIND again.
The city name is always followed by the word “SALARY”. So, if we search for the position of that word and subtract the position of “BORN,” we’ll get the length of what’s between the two. The ultimate formula looks like spaghetti but works just fine:
Used on the example text, it returns:
Dunedin, FL Lexington, KY St. Louis, MO Miami, FL Hialeah, FL Robstown, TX Sacramento, CA Coldwater, OH Indianapolis, IN Auburndale, WI
That’s it. Fairly handy and further proof that Excel is a versatile part of the data journalists’ tool kit.