Valid email address:

[SASPhile]

As a first step towards data cleaning:

How to get the counts of following where there just hypens before @ in email addresses:

------------------------------------@GMAIL.COM

-------------@-------------.COM

-------------@YAHOO.COM

------------@EMC.COM

----------@HOTMAIL.COM

----------@YAHOO.COM

---------@PRTC.NET

---------@--.COM

--------@AOL.COM

abc@gmail.com

[Haikuo]

You will probably need some Regular expression:

data have;

  input email :$char100.;

  cards;

  ------------------------------------@GMAIL.COM

-------------@-------------.COM

-------------@YAHOO.COM

------------@EMC.COM

----------@HOTMAIL.COM

----------@YAHOO.COM

---------@PRTC.NET

---------@--.COM

--------@AOL.COM

abc@gmail.com

mike-smith@tg-aol.com

;

data want;

set have;

  if prxmatch("/^-+@/",email) then ind=1;

run;

proc print;run;

Haikuo

[Astounding]

One way:

proc freq data=have;

tables email;

where index(email, '----');

run;

It will print a table based on any email addresses that contain four dashes in a row, anywhere within the address.  If you want to cut it to less than 4, that's up to you.  You know your data best.

Good luck.

[Haikuo]

I had to ask, what if you want to keep:

abc----xyz@gmail.com

but to drop:

----@gmail.com?

I still think the Regular expression is the way to go, at least for the first step.

my 2 cents,

Haikuo

[Astounding]

I suppose I could switch from this:

index(email, '----')

to this:

index(email, '-@')

[art297]

: Since you wanted strings that start with any number of hyphens, then followed with an @, I'll suggest adding one more character to Haikuo's code.  The ^ at the beginning forces the match to have occurred at the start of the string:

data have;

  informat email $50.;

  input email;

  cards;

------------------------------------@GMAIL.COM

-------------@-------------.COM

-------------@YAHOO.COM

------------@EMC.COM

x----------@HOTMAIL.COM

----------@YAHOO.COM

---------@PRTC.NET

---------@--.COM

--------@AOL.COM

abc@gmail.com

;

data want;

  set have;;

  if prxmatch("/^-+@/",email) then ind=1;

run;

[DBailey]

DATA WANT;

SET HAVE;

if index(email,'@')-1 = countc(substr(email,1,index(email,'@')-1),'-');

run;