topic Re: Address cleaning in SAS Procedures

Address cleaning

huongc2 — Fri, 20 Nov 2020 15:11:10 GMT

Hi everyone,

I am having some troubles with cleaning these addresses. at first it was easy with some if statement. But since my data is getting bigger, it is not efficient to use if statement anymore. I would love to hear your thoughts about how to clean this field.

I have:

Data have;

input address $200;

datalines;

100 ABC ST RM S02A

102 ABC STREET FLOOR 1

103 ABC ST APT 3 HOMELESS

1035 CD AVENUE FLOOR 2

108 SOMETHING ST # 2FL

115 VISA VISTA DR APT 212 APT 212

1155 LOOK AVENUE APT 205

12 BORED AVE APT 2

1214 TIRED STREET APT 428

127 HAPPY STREET FLOOR 2

1397 SOMEWHERE STREET FIRST FLOOR

142 SOMETHING ST APT 3

200 RAINBOW AVE UNIT 202

;

I don't want any Unit or floor or apt number in the clean address. So I want the address field that would look like this:

100 ABC ST

102 ABC STREET

103 ABC ST

1035 CD AVENUE

108 SOMETHING ST

115 VISA VISTA DR

1155 LOOK AVENUE

12 BORED AVE

1214 TIRED STREET

127 HAPPY STREET

1397 SOMEWHERE STREET

142 SOMETHING ST

200 RAINBOW AVE

Thank you so much!

Re: Address cleaning

SASKiwi — Fri, 20 Nov 2020 22:22:30 GMT

What is your actual use case? If your address cleaning needs to scale up to enterprise-wide customer volumes and techniques then you would be better off using a tool specific to this task like SAS Data Quality. On the other hand, cleaning a few hundred addresses with a few transformation rules like the ones in your post, you are probably better off persevering with your current approach.

Re: Address cleaning

huongc2 — Mon, 23 Nov 2020 14:16:24 GMT

Ah Thank you. I thought SAS would have function for this that I don't know 😄