Separate Column Using SED.

amy441 · 06-11-2018, 03:18 AM

Yes guys, but I'm stuck that's why I'm asking for a little help. If you give me some hit then that would be appreciated.

Also, I'm able to solve all other assignments my own but just stuck with this one. :/

pan64 · 06-11-2018, 05:44 AM

I gave you some tips, you need to brake it down into small steps and implement them one by one. You can also simplify the task first.

scasey · 06-11-2018, 09:41 AM

A significant part of the challenge of the assignment to split city,state into two separate fields is that not all of the lines have fields containing city also include a state. (This is one of the reasons sed may not be the best solution)

pan64 has outlined a "plan of attack" that should work, but again, it's your work to do.

rtmistler · 06-11-2018, 10:06 AM

Is this the entire body of input text you need to contend with?

Quote:

Originally Posted by amy441

This is an output of cat abc.csv file command. This is actually a homework given by my teacher and he specifically mentioned to split the column using SED.

Year,Rank,Organization ID,Organization Name,Organization Location,Private Income,Total Assets,Service Expense,Fundraising Expense
2004,1,321092,Salvation Army,"Alexandria, Va.","$1,324,089,000","$3,039,922,000","$2,126,200,000","$123,791,000"
2004,2,321148,American Cancer Society,Atlanta,"$794,000,000","$836,295,000","$610,639,000","$188,150,000"
2004,3,321036,Gifts In Kind International,"Alexandria, Va.","$787,192,199","$790,561,819","$792,432,766","$167,242"

I feel you should consider visiting with your instructor and asking them what the full point of this assignment truly is.

You can edit a file using sed. You can also use tr, and awk. Each tool has benefits and some drawbacks.

Well understood that they told you to only use sed.

A HUGE concern for me is summarized here:

Quote:

,"Alexandria, Va.",
,Atlanta,
,"Alexandria, Va.",

They've been inconsistent twice with their text for this problem.

That middle term is not enclosed in quotations

It also does not contain a state.

If their intention was to have you stretch your mind to properly change the first and third terms solely using sed, that's one thing.

Is the whole point of their assignment to be able to contend with the first and third terms, as well as the second term?

Is there more input text you haven't shown by the way?

Are you supposed to fix the quotations around "Alexandria, VA" so that it becomes "Alexandria", "VA",?

Are you supposed to add quotations around Atlanta?

And as asked by another person, what are you supposed to do about that term missing the state?

allend · 06-11-2018, 10:52 PM

I agree that sed is a poor choice of tool for this exercise. I would be looking at gawk.

syg00 · 06-11-2018, 11:14 PM

Missing fields are simply consecutive commas in a CSV. It's not actually that hard to do the entire exercise in sed - doing it in one stanza gets interesting - but it is an exercise for the OP.

pan64 · 06-12-2018, 01:49 AM

in general a huge list of seds like this:

Code:

s/"Alexandria, Va.",/,"Alexandria", "Va."/

may work.
Who knows?

scasey · 06-12-2018, 01:42 PM

I've spent way more time in the bowels of csv files than I ever wanted to.

Just to clarify a couple of things:

The fields are separated by commas.
If the field contains a comma, then it is quoted.

So, in the second row, the lack of a comma and state is why that field is not quote

If the OP could save the data from the spreadsheet in a tab-delimited format, that would remove all the delimiting commas and the quotes from the data, which might help things some.
In any case, it's important to differentiate between commas that are delimiters and commas that are data, and why the quotes are there.