Separate Out STOCK Symbols From Large Text File

[nkormanik]

Assume:  Stock symbols are composed of CAPITAL LETTERS.

What might be an efficient way of extracting out all stock symbols from a large text file, with stock symbols appearing here and there throughout?

Thanks much!

Nicholas Kormanik

[skwire]

Can you provide us with the text file?  Or a sample thereof?

[nkormanik]

Here is an example of part of such a text file.  I am not expecting perfection, mind you.  Some further editing will likely be required.  Again, let's assume that NOTHING other than caps will be allowed, so excluding parentheses, colons, commas, numbers, etc.:

[Select]

(FWONA), Facebook (FB), Universal Display (OLED), Apple (AAPL), SBA Communications (SBAC), and Abiomed (ABMD

The daily volume for an average stock on the S&P 500 is 4 million shares, while the median is slightly more than 2 million. Bank of America Corp’s BAC, -0.29% average daily volume is about 87 million shares, followed by Apple Inc’s AAPL, +0.37% at roughly 54 million. By comparison less than 300,000 shares of Dun & Bradstreet Corp. DNB, +0.28%  changed hands in the average day.
Other standouts include energy, airline and biotechnology stocks - small and large. Since the start of the year, on average, about 20 million shares of Chesapeake Energy CHK, -1.81% changed hands daily, (The stock lost more than half its value since the start of the year.) Halliburton Co. HAL, +0.57% Exxon Mobil XOM, +0.51% Transocean Ltd. RIG, +3.51% Delta Air Lines DAL, +3.25% American Airlines Group AAL, +5.43% General Motors GM, +0.23% and Ford Motors F, +1.03% saw people churn their stocks.

Symbol Company Last Sale* Change Net / % Share Volume
AAPL
Apple Inc. Apple Inc. $ 159.86 0.59 ? 0.37% 24,783,937
QQQ
PowerShares QQQ Trust, Series 1 $ 141.97 0.30 ? 0.21% 24,236,348
TVIX
VelocityShares Daily 2x VIX Short-Term ETN $ 16.90 1 ? 5.59% 22,703,370
CSCO
Cisco Systems, Inc. Cisco Systems, Inc. $ 31.44 0.20 ? 0.64% 18,206,334
JD
JD.com, Inc. JD.com, Inc. $ 40.87 1.20 ? 2.85% 15,610,100
MOMO
Momo Inc. $ 35.59 2.39 ? 6.29% 15,112,039
BRCD

[nkormanik]

If someone here wants to actually create a program to do the above, perhaps it might be called:  Stock Symbols Only.

(Skwire's Stock Symbols Only??)

I think what I am asking for can be accomplished with RegEx, but not sure.

A little Windows program, even a command-line program, asking for the text file input, and spitting out results to a separate file would be perfect.  Prefer symbols to be one per line, but not necessary, and no need to sort.

[skwire]

So, from that example, you just want a list like this?:

[Select]

FWONA
FB
OLED
AAPL
SBAC
ABMD
BAC
DNB
CHK
HAL
XOM
RIG
DAL
AAL
GM
F
QQQ
TVIX
ETN
CSCO
JD
MOMO
BRCD

[nkormanik]

Wow.  Ohhh my goodness.  How'd you do that??!!

Exactly!

[skwire]

I just created that manually as an example output.

[skwire]

As you mentioned in your first post, this sort of thing is a bit inexact and might require a bit of manual cleanup.  However, please try the attached EXE as it should be pretty close to what you want.

[nkormanik]

I placed your sso.exe and a text file, test.txt, into C:\2 and executed, dragging test.txt onto the open box appearing.  Named output file symbols.txt:

Error!
There was a problem saving the output file.  Perhaps it's open in another application.

I tried a number of other attempts, with same error message.

[Shades]

Are you sure that Windows allows you to write in that folder. Run the exe with 'Run as administrator' option.

If you don't want to do that, you better place the executable in your user folder and try again.

[nkormanik]

Same error.  Something is apparently wrong with my system.  Bummer.

[skwire]

Same error.  Something is apparently wrong with my system.  Bummer.

-nkormanik (August 27, 2017, 09:09 AM)

Nah, it was my fault.  Apologies.     Please redownload and try again.

[nkormanik]

Beautiful!  Works exactly as I was hoping!

Hope others out there might find and use this as well.

I'm renaming my personal copy: Skwire's Stock Symbols Only.

Thanks a million, Skwire.

[4wd]

@skwire: What regex did you use, (if you did)?

I had a quick look at making a single line PoSh command but my brain wasn't up to the nuances of stopping it grabbing every uppercase letter.  regardless of what followed

[Ath]

Well, I tried a regex too, but I couldn't fix 1 issue.
My regex:

[Select]

\b[A-Z]\b

But it keeps grabbing both the S and P from S&P, and I didn't want to replace the \b with a (rather complex) group construction of all possible punctuation to solve this so I stopped searching.

[skwire]

@skwire: What regex did you use, (if you did)?

-4wd (August 28, 2017, 12:19 AM)

What ended up working the best was:

1. Loop through each line.
2. Crack each line on its spaces.
3. Evaluate each part with: [A-Z]+\b

It's not quite perfect, and I'm not sure it ever could be, but it seems to be close enough for the OP's work.

[4wd]

I was trying to end up with something along the lines of:

Code: PowerShell [Select]

1. (Get-Content -Path .\test.txt) -creplace "\b[^A-Z]+\b" " " | Set-Content -Path out.txt

Would have left spaces between the remaining terms to make output formatting easier ... wasn't getting the RegEx though.

That was the theory anyway 

Now I think about it I could have split the input easily:

Code: PowerShell [Select]

1. [regex]::Split((Get-Content -Path .\test.txt),'[\s,\.\(\)]') | %{ if($_ -cmatch('^[A-Z]+$')){Write-Host $_}}

Output is OK except for outputting SBA and SBAC and duplicated entries.

@skwire: How did you stop SBA being passed through since SBAC is the stock code, (considering codes occur both inside and outside of surrounding brackets)?

You must be checking for duplicated entries also since there are multiple occurrences of JD and QQQ.

Was something to blow a few cobwebs out of my head anyway 

Addendum:
Duplicates removed courtesy of http://www.secretgeek.net/ps_duplicates

Code: PowerShell [Select]

1. $hash = @{};[regex]::Split((Get-Content -Path .\test.txt),'[\s,\.\(\)]') | %{if($_ -cmatch('^[A-Z]+$')){if($hash.$_ -eq $null) { $_ }; $hash.$_ = 1}} > .\output.txt

Output

FWONA
FB
OLED
AAPL
SBA
SBAC
ABMD
BAC
DNB
CHK
HAL
XOM
RIG
DAL
AAL
GM
F
QQQ
TVIX
VIX
ETN
CSCO
JD
MOMO
BRCD

[skwire]

You must be checking for duplicated entries also since there are multiple occurrences of JD and QQQ.

-4wd (August 28, 2017, 08:23 PM)

Yes, at the end of the list processing, I do an alpha sort and filter out duplicates.