Remove common http pattern

As URLs are beginning with http://, so I removed this part by a [...]]]> Recently, I got a tough task about collecting external sources link back to my site. It may be a piece cake to most of the people, but what I’ve got is a list of URL and an excel.

Remove common http pattern

As URLs are beginning with http://, so I removed this part by a simple formula:

RIGHT(,LEN()-7) OR RIGHT(,LEN()-LEN(“http://”))

Since http:// is 7 characters length, that’s why len()-7 or len()-len(“http://”) both return the remaining number of characters left.

Remove common sub-domain pattern (www)

Another part to trim at the front of the domain name is another common pattern – www. For those domain name not starting with this pattern, such as blog.joe-ho.com, I’ll handle it later.

IF(LEFT(,4)=”www.”,,RIGHT(,LEN()-4))

Actually this is similar as the previous step we did, but added an IF condition.

We can combine both simple formula into a complex way, when the domain is starting with “www” we take out the first 11 characters (“http://www.”), otherwise we just take out http header. The complex formula will be same as below:

IF(MID(,8,4)=”www.”,RIGHT(LEN(),LEN()-11),RIGHT(LEN(),LEN()-7))

Remove remaining after domain name

Next step we will focus on cut out the remaining after domain names. As some URLs are shown detaily which pages linked with our websites, those non-standard URLs still have a same pattern starting with a / mark. Our next mission is removing all the stuffs after that mark, sure we will remove the mark too. We use the following formula to find out where the mark located in cell.

LEFT(,FIND(“/”,)-1)

The FIND forumla is doing a straight forward task, find the location where the text found in a cell. If nothing found, the forumla will return 0. That mean when forumla return more than 0, ie >0, that mean the text is found in cell.

After the FIND forumla, I minus 1 in the result. It is because we need remove the “/” mark too. If we didn’t do so, the final domain name will show as joe-ho.com/ instead of joe-ho.com. Fianlly, we use LEFT forumla to get the most valuable things we needed.

If the text we need happened more than 1 time, we can use SUBTITUTE formula to change the text to a special symbol and use FIND forumla to locate that symbol. Below is the example of the formula, remember we need always specific where the mark we need.

LEFT(,FIND(” “,SUBSTITUTE(,”/”,” “,))-1)

Hope the above is useful for you. Discusses / comments are welcome.

下次有機會再同各位研究一下.