While working on a project recently, I needed a simple list of US State names. My goal was to have a plain text file, one state per line, no funny business. The order of the states wasn't important. I always like a command-line challenge, so my first thought was "find the list online somewhere, wget or curl as needed, slice up the html, and be done."1
So, I went to the source of all information, and lo! there is a table whose first column is all the US states. #fistpump The only problem is I really wasn't planning to parse HTML for this little exercise; seems like a bit of overkill. But then I noted the [edit] link next to the table name...
... and clicked through that to observe just the table source. Now this we can work with!
I simply copy/pasted this plain text into a file locally (say, states-table.txt). Then, had a look:
$ head -n30 states-table.txt
==States==
{| class="wikitable sortable plainrowheaders" style="text-align: right;"
|+ States of the United States
!scope="col" width=12%|State
!scope="col" width=1%|[[List of U.S. state abbreviations|Abbr.]]
!scope="col" width=8%|[[Capital city|Capital]]
!scope="col" width=12%|Largest city{{#tag:ref|The largest city is the city in a state with the largest population in the [[city proper]]; [[metropolitan area]]s are not considered in this number.|group=upper-alpha|name=a}}<ref name="State and Local Government Finances and Employment">{{cite web|url=http://www.census.gov/compendia/statab/2012/tables/12s0448.pdf|title=State and Local Government Finances and Employment|year=2012|publisher=[[United States Census Bureau]]|pages=284|type=pdf|accessdate=July 8, 2013}}</ref>
!scope="col" width=11% data-sort-type="date"|Statehood
!scope="col" width="12%"|Population<br>(2013 est)<ref name=PopEstUS>{{cite web|url=http://www.census.gov/popest/data/state/totals/2013/tables/NST-EST2013-01.xls|title=Table 1. Annual Estimates of the Population for the United States, Regions, States, and Puerto Rico: April 1, 2010 to July 1, 2013|format=[[Microsoft Excel]]|work=2013 Population Estimates|publisher=[[United States Census Bureau]], Population Division|date=December 2013|accessdate=January 12, 2014}}</ref>
!scope="col" width=12%|Total area in mi<sup>2</sup> (km<sup>2</sup>){{#tag:ref|Area figures are rounded to the nearest whole number.|group=upper-alpha|name=b}}<ref name="2010 Census of Population and Housing">{{cite web|url=http://www.census.gov/prod/cen2010/cph-2-1.pdf|title=2010 Census of Population and Housing|date=September 2012|publisher=[[United States Census Bureau]]|pages=41|type=pdf|accessdate=July 8, 2013}}</ref>
!scope="col" width=12%|Land area in mi<sup>2</sup> (km<sup>2</sup>)<ref group=upper-alpha name=b /><ref name="2010 Census of Population and Housing" />
!scope="col" width=12%|Water area in mi<sup>2</sup> (km<sup>2</sup>)<ref group=upper-alpha name=b /><ref name="2010 Census of Population and Housing" />
!scope="col" width=1%|[[List of United States congressional districts|House seat(s)]]
|-
!scope="row"|{{flag|Alabama}}
|AL
|[[Montgomery, Alabama|Montgomery]]
|[[Birmingham, Alabama|Birmingham]]
|December 14, 1819
|4,833,722
|{{Convert|52420|mi2|km2|sigfig=6|abbr=values|sortable=on}}
|{{Convert|50645|mi2|km2|sigfig=6|abbr=values|sortable=on}}
|{{Convert|1775|mi2|km2|sigfig=4|abbr=values|sortable=on}}
|7
|-
!scope="row"|{{flag|Alaska}}
|AK
|[[Juneau, Alaska|Juneau]]
|[[Anchorage, Alaska|Anchorage]]
|January 3, 1959
Clearly there's a lot of extra formatting for the other columns, but we can see a couple of the state names, so we can definitely observe some patterns. A simple grep and sed combination get us almost to the finish line:
$ cat states-table.txt | grep "flag|" | sed 's/^.*flag|//;s/}}.*//'
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia (U.S. state)|name=Georgia
Hawaii
...
One extra bit of sed formatting (strip anything that's not a letter)...
$ cat states-table.txt | grep "flag|" | sed 's/^.*flag|//;s/}}.*//;s/ [^a-Z].*$//'
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
...
... and we're good to go! (note: if you're doing this on OS X, you may have to swap that last sed range for [^a-zA-Z])
It's always nice to flex your command-line fu.
comments powered by Disqus