import from CSV into Ruby array, with 1st field as hash key, then lookup a field's value given header row

Question

Maybe somebody can help me.

Starting with a CSV file like so:

Ticker,"Price","Market Cap"
ZUMZ,30.00,933.90
XTEX,16.02,811.57
AAC,9.83,80.02

I manage to read them into an array:

require 'csv'
tickers = CSV.read("stocks.csv", {:headers => true, :return_headers => true, :header_converters => :symbol, :converters => :all} )

To verify data, this works:

puts tickers[1][:ticker]
ZUMZ

However this doesn't:

puts tickers[:ticker => "XTEX"][:price]

How would I go about turning this array into a hash using the ticker field as unique key, such that I could easily look up any other field associatively as defined in line 1 of the input? Dealing with many more columns and rows.

Much appreciated!

Answer 1

To get the best of both worlds (very fast reading from a huge file AND the benefits of a native Ruby CSV object) my code had since evolved into this method:

$stock="XTEX"
csv_data = CSV.parse IO.read(%`|sed -n "1p; /^#{$stock},/p" stocks.csv`), {:headers => true, :return_headers => false, :header_converters => :symbol, :converters => :all}

# Now the 1-row CSV object is ready for use, eg:
$company = csv_data[:company][0]
$volatility_month = csv_data[:volatility_month][0].to_f
$sector = csv_data[:sector][0]
$industry = csv_data[:industry][0]
$rsi14d = csv_data[:relative_strength_index_14][0].to_f

which is closer to my original method, but only reads in one record plus line 1 of the input csv file containing the headers. The inline sed instructions take care of that--and the whole thing is noticably instant. This this is better than last because now I can access all the fields from Ruby, and associatively, not caring about column numbers anymore as was the case with awk.

Answer 2

Like this (it works with other CSVs too, not just the one you specified):

require 'csv'

tickers = {}

CSV.foreach("stocks.csv", :headers => true, :header_converters => :symbol, :converters => :all) do |row|
  tickers[row.fields[0]] = Hash[row.headers[1..-1].zip(row.fields[1..-1])]
end

Result:

{"ZUMZ"=>{:price=>30.0, :market_cap=>933.9}, "XTEX"=>{:price=>16.02, :market_cap=>811.57}, "AAC"=>{:price=>9.83, :market_cap=>80.02}}

You can access elements in this data structure like this:

puts tickers["XTEX"][:price] #=> 16.02

Edit (according to comment): For selecting elements, you can do something like

 tickers.select { |ticker, vals| vals[:price] > 10.0 }

Answer 3

CSV.read(file_path, headers:true, header_converters: :symbol, converters: :all).collect do |row|
  Hash[row.collect { |c,r| [c,r] }]
end

Answer 4

---

To add on to Michael Kohl's answer, if you want to access the elements in the following manner

puts tickers[:price]["XTEX"] #=> 16.02

You can try the following code snippet:

CSV.foreach("Workbook1.csv", :headers => true, :header_converters => :symbol, :converters => :all) do |row|
    hash_row =  row.headers[1..-1].zip( (Array.new(row.fields.length-1, row.fields[0]).zip(row.fields[1..-1])) ).to_h
    hash_row.each{|key, value| tickers[key] ? tickers[key].merge!([value].to_h) : tickers[key] = [value].to_h}
end

Answer 5

Not as 1-liner-ie but this was more clear to me.

csv_headers = CSV.parse(STDIN.gets)
csv = CSV.new(STDIN)

kick_list = []
csv.each_with_index do |row, i|
  row_hash = {}
  row.each_with_index do |field, j|
    row_hash[csv_headers[0][j]] = field
  end
  kick_list << row_hash
end

Answer 6

While this isn't a 100% native Ruby solution to the original question, should others stumble here and wonder what awk call I wound up using for now, here it is:

$dividend_yield = IO.readlines("|awk -F, '$1==\"#{$stock}\" {print $9}' datafile.csv")[0].to_f

where $stock is the variable I had previously assigned to a company's ticker symbol (the wannabe key field). Conveniently survives problems by returning 0.0 if: ticker or file or field #9 not found/empty, or if value cannot be typecasted to a float. So any trailing '%' in my case gets nicely truncated.

Note that at this point one could easily add more filters within awk to have IO.readlines return a 1-dim array of output lines from the smaller resulting CSV, eg.

 awk -F, '$9 >= 2.01  &&  $2 > 99.99  {print $0}' datafile.csv 

outputs in bash which lines have a DivYld (col 9) over 2.01 and price (col 2) over 99.99. (Unfortunately I'm not using the header row to to determine field numbers, which is where I was ultimately hoping for some searchable associative Ruby array.)