Print The Formula

Question

Introduction

In chemistry there is a type of extension, .xyz extension,(https://en.wikipedia.org/wiki/XYZ_file_format), that prints in each line a chemical element, and the coordinates in the plane of the element. This is very useful for chemists to understand chemical compounds and to visualize the compounds in 3D. I thought it would be fun to, given a .xyz file, print the chemical formula.

Challenge

Given an .xyz file, print the chemical formula of the compound in any programming language in the smallest possible number of bytes. Note:

• Originally, the input was to be given as a file. As I have been pointed out, this constraints the challenge. Therefore you may assume the input is a list/array of strings, each representing a line of the .xyz file.
• There are no restrictions in the ordering of the elements.
• Each element should be printed with an underscore "_" delimiting the element and the number of times it appears
• The first two lines of any .xyz file is the number of elements, and a comment line (keep that in mind).

Example Input and Output

Suppose you have a file p.xyz which contains the following (where the first line is the number of elements, and the second a comment), input:

5  
A mystery chemical formula...  
Ba      0.000   0.000  0.000  
Hf      0.5     0.5    0.5  
O       0.5     0.5    0.000  
O       0.5     0.000  0.5  
O       0.000   0.5    0.5  

---

Output:
Ba_1Hf_1O_3

---

Testing

A quick test is with the example mentioned. A more thorough test is the following: since the test file is thousands of lines, I'll share the .xyz file:
https://gist.github.com/nachonavarro/1e95cb8bbbc644af3c44

Answer 1

Pyth - 18 bytes

sjL\__MrShMcR;ttQ8

Try it online here.

Answer 2

Japt, 21 bytes

U=¢m¸mg)â £X+'_+Uè_¥X

Test it online! Input is given as an array of strings (which can be formatted as in the link).

Ungolfed and explanation

U=¢   m¸  mg)â £    X+'_+Uè_  ¥ X
U=Us2 mqS mg)â mXYZ{X+'_+UèZ{Z==X

          // Implicit: U = input array of strings
Us2       // Slice off the first two items of U.
mqS mg    // Map each item by splitting at spaces, then taking the first item.
U=    )   // Set U to the result.
â mXYZ{   // Uniquify, then map each item X to:
UèZ{Z==X  //  Count the number of items Z in U where Z == X.
X+'_+     //  Prepend X and an underscore.
          // Implicit output

Answer 3

AWK, 44

NR>2{a[$1]++}END{for(i in a)printf i"_"a[i]}

Try it online.

Answer 4

Shell + GNU Utilities, 67

sed '1d;2d;s/ .*//'|sort|uniq -c|sed -Ez 's/\s*(\S+) (\S+)/\2_\1/g'

Try it online.

Answer 5

Mathematica, 79 53 bytes

StringRiffle[Tally@StringExtract[#[[3;;]],1],"","_"]&

Quite simple.