SED to extract HTML text data, not quite right!

Shell Programming and Scripting
1

I am attempting to extract weather data from the following website, but for the Victoria area only:

Text Forecasts - Environment Canada

I use this:

 sed -n "/Greater Victoria./,/Fraser Valley./p"

But that phrasing does not sometimes get it all and think perhaps the website has more than one linefeed, carriage return, whatever, that messes up my coding. Any ideas appreciated.

Larry

2

Hi.

If I copy the text from the site, and modify your statement slightly:

sed -n "/Greater Victoria./,/^$/p" file1

I get this output:

Greater Victoria.
Monday..Cloudy. High 8.
Tuesday..Cloudy with 60 percent chance of showers. Low plus 3.
 High 7.
Wednesday..Cloudy with 40 percent chance of flurries or rain showers.
 Low plus 3. High 8.
Thursday..Cloudy with 60 percent chance of rain showers mixed with
 flurries. Low plus 3. High 8.
Friday..Cloudy with 30 percent chance of showers. Low plus 3. High 9.
Normals for the period..Low plus 2. High 7.

It looks exactly the same as on the website. Is it not want you wanted?

3

Google weather api. There's a bunch of them.

For example, I looked at Yahoo! Weather RSS Feed to come up with this for Victoria weather:

curl -s "http://weather.yahooapis.com/forecastrss?w=9848&u=c"|sed "1,/Current Conditions/d"|head -n1

The text is simpler and more predictable than a regular web site, so it shouldn't be too difficult to parse out the interesting items.