Re: delete columns csv file

On 4/4/2008 12:32 AM, Hermann Peifer wrote:
> Ed Morton wrote:
> 
>>On 4/3/2008 3:27 PM, Hermann Peifer wrote:
>>
>>>Ed Morton wrote:
>>>
>>>
>>>>On 4/3/2008 2:19 PM, r wrote:
>>>>
>>>>
>>>>>Readers,
>>>>>
>>>>>I have a csv file (greater than 256 columns hence unable to open and
>>>>>edit in
>>>>>spreadsheet) of the following format:
>>>>>
>>>>>column header1, column header2, column header3
>>>>>1,0.0e0,0.0e0,5e-6
>>>>>2,0.0e0,0.0e0,6e-7
>>>>>3,0.0e0,0.0e0,0.0e0
>>>>>
>>>>>I want to perform: "if column headerx contains only values of 0.0e0,
>>>>>delete the column (including the column header).
>>>>>
>>>>>Can a script be provided and then a discussion of how this problem is
>>>>>solved please (novice)?
>>>>>
>>>>>Yours,
>>>>>
>>>>>mandriva 2008
>>>>
>>>>You need to parse the file twice:
>>>>
>>>>awk -F, '
>>>>NR==FNR {
>>>>  if (NR>1)
>>>>     for (i=1;i<=NF;i++)
>>>>        if ($i"" != "0.0e0")
>>>>           good[i]
>>>>  next
>>>>}
>>>>{  out=sep=""
>>>>  for (i=1;i<=NF;i++) {
>>>>      if (i in good)
>>>>         out = out sep $i
>>>>      sep=FS
>>>>  } print out
>>>>}' file file
>>>>
>>>>Regards,
>>>>
>>>>	Ed.
>>>>
>>>
>>>Shouldn't "if column headerx contains only values of..." result into:
>>>
>>>if (NR==1)
>>>
>>
>>That's what I thought at first too, but then I reread it and I think it
means
>>"if all values in a given column are ....".
>>
>>	Ed.
>>
> 
> 
> OK. Then it depends on what the OP really meant.
> 
> Couldn't: if ($i"" != "0.0e0") be replaced by a simple: if ($i)  ?

No, because the OP asked specifically for a comparison to "0.0e0" not for
"0" or
"". It probably could've been just ($i != "0.0e0") but I couldn't be
bothered
thinking about the string/numeric comparison rules...

> And I wasn't sure what the first pair of double quotes in this condition

> was meant to do?

Just ensuring a string comparison by making both sides of the comparison
strings.

	Ed.