Twitter API Question posted in
The official Twitter API documentation can be found here.

Tee pipe into 3 different processes and grepping the second match – Twitter API

I am trying to create a bash script which shows me the latest stats about corona infection numbers in the countries Germany and Switzerland and also in the whole world.

corona () {
    curl -s https://corona-stats.online?minimal=true | tee >(head -n 1) > >(grep "(CH)|(DE)")
    curl -s https://corona-stats.online?minimal=true | tail -n 20 | grep World
}

As you can see, to do this I had to create this very ugly script where curl is called twice. I had to do this because the website looks like this:

Rank World                               Total Cases  New Cases ▲ Total Deaths New Deaths ▲ Recovered  Active    Critical Cases / 1M pop
1    USA (US)                               7,497,256     2,585 ▲      212,694         34 ▲  4,737,369 2,547,193   14,190         22,617
2    India (IN)                             6,397,896     5,936 ▲       99,833         29 ▲  5,352,078   945,985    8,944          4,625
3    Brazil (BR)                            4,849,229                  144,767               4,212,772   491,690    8,318         22,773
4    Russia (RU)                            1,194,643     9,412 ▲       21,077        186 ▲    970,296   203,270    2,300          8,185
...
22   Germany (DE)                             295,943       413 ▲        9,586                 259,500    26,857      362          3,529
...
58   Switzerland (CH)                          54,384       552 ▲        2,075          1 ▲     45,300     7,009       32          6,272
...
     World                                 34,534,040    63,822 ▲    1,028,540      1,395 ▲ 25,482,492 8,023,008   66,092       4,430.85


Code: https://github.com/sagarkarira/coronavirus-tracker-cli
Twitter: https://twitter.com/ekrysis

Last Updated on: 02-Oct-2020 12:10 UTC

US STATES API: https://corona-stats.online/states/us
HELP: https://corona-stats.online/help
SPONSORED BY: ZEIT NOW
Checkout fun new side project I am working on: https://messagink.com/story/5eefb79b77193090dd29d3ce/global-response-to-coronavirus

I only want to display the first line, the last line of the table (World) and the two lines about Germany and Switzerland. I manged to display the first line as well as the two countries by piping the output of curl into head -n 1 and grepping the country codes. I was able to do both things thanks to this answer.

Now I want to get the last line in the table, the one where the current cases of the whole World are displayed. I tried to use tee again to pipe it into a third process tee >(head -n 1) > >(grep "(CH)|(DE)") > >(tail -n 20 | grep World). But that didn’t work. My first question is, how can I pipe an output into 3 different processes using tee?

The second question revolves around the way I try to grep the World line. I tail the last 20 lines and then grep "World". I do this because if I simply grep "World", it only return the title line where "World" can also be found. So my second question is: How can I grep only the last or second occurance?

Tags:

3

Answers


  1. 0

    You can chain several tee commands and throw away only the last output of tee:

    curl -s ... | tee >( cmd1 ) | tee >( cmd2 ) | tee > >( cmd3 )

    Actually, we can shorten it to:

    curl -s ... | tee >( cmd1 ) | tee >( cmd2 ) | cmd3

    because we do not use the output of the last tee anyway.

    Having multiple commands write to the terminal at the same time might get the output mixed up. A much more elegant solution is to use only one grep, e.g.

    curl -s ... | grep '(DE)|(CH)|World.*,'

    The expression World.*, will just look for a comma in the same line after World, in order to exclude the head line.

    Login or Signup to reply.
  2. 0

    You can try this:

    curl -s https://corona-stats.online?minimal=true | grep -E "(Rank|^1[^0-9]|(CH)|(DE))"

    Use grep to display only line contain "Rank", 1[non-digit], (CH), (DE)

    Login or Signup to reply.
  3. 0

    I think a variable should suit better what you need (at least in this case), something like:

    corona() {
  data="$(curl -s https://corona-stats.online?minimal=true)"
  echo "$data" | head -n 1
  echo "$data" | grep "(CH)|(DE)"
  echo "$data" | tail -n 20 | grep World
}

    It would convey easier what you’re trying to do and would also be easier to expand if you’d need to change anything.

    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search