|
View previous topic :: View next topic
|
| Author |
Message |
Roy Ware
New User
Joined: 19 Sep 2013
Posts: 14
Location: USA
|
|
|
|
I'm being held hostage by two database tables who refuse to compromise. One table has the City Name in Mixed Case letters: El Paso, for example. The second has the City Name in Uppercase letters: EL PASO.
The goal is to summarize the dollar amount field for each city, regardless of the case, with the mixed case being the desirable output, but showing the uppercase name in case a "matching" lowercase is not available.
Here's the data:
| Code: |
Mattoon 277.86
MATTOON 179.02
Charleston 801.13
POPLAR GROVE 354.03 |
The desired output is:
| Code: |
Mattoon 456.88
Charleston 801.13
POPLAR GROVE 354.03 |
Of course, nothing like this has been attempted here before, so there aren't any examples and the SyncSort manual isn't much help. Any suggestions? |
|
| Back to top |
|
 |
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10888
Location: italy
|
|
|
|
| just overlay/build an additional field with the content of the key upper/lower cased for matching |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
As enrico said, "normalise" the name in a temporary extension. SORT on that name. You'll also need to discriminate between upper/mixed, so that you can favour the mixed when both are present (a second field to include an indicator for mixed/upper so you can "outsort" the upper).
Exactly how you do that depends on the quality of your data. If you can have embedded typos...?
Whether SyncSort has TRAN= or some other built-in way to convert case is another issue. |
|
| Back to top |
|
|
Roy Ware
New User
Joined: 19 Sep 2013
Posts: 14
Location: USA
|
|
|
|
| Yes, the field can have embedded typos - why can't we hold the front-ends to the same standards we have to match? - but Enrico's suggestion is going to work great! |
|
| Back to top |
|
|
|
|
