About brulard

brulard · ‎03-22-2017

thank you very much Laurie for your suggestion and for some of the explanations provided.

brulard · ‎03-21-2017

Thanks LaurieF for ur reply. At it's simplest, I need to randomly select 5 observations per agent_name, that should follow this pattern: - if can find 3 observations of risk_level=1, flag subject_to_review=Y -if can 2 observations of risk_level=2, flag subject_to_review=Y This would arrive at 5 observations per agent_name. it becomes more complicated if cannot find 3 observations, as then need to pick up a lower risk level to arrive at your five per agent (no hard rule, but should first fetch higher risk level. From higher to lower is : 1,2,3,0,'') See below example, but without it being randomly assigned Y. If no one has a lead to suggest, I may go with attempting to achieve this with excel vba. Thank you DATA_WANT Agent_name action_type risk_level subject_to_review Albert 1000 1 Y Albert 1000 1 Y Albert 1000 1 Albert 1000 1 Albert 2100 2 Y Albert 2100 2 Y Albert 2100 2 Y Albert 2100 2 Albert 3230 0 Albert 2500 3 Albert 2500 3 Albert 2500 3 Albert 2500 3 Albert 2500 3 Albert 5200 Barb 1000 1 Y Barb 1000 1 Y Barb 2100 2 Y Barb 2100 2 Y Barb 2100 2 Y Barb 3230 0 Cathy 1000 1 Y Cathy 2100 2 Y Cathy 2100 2 Y Cathy 3230 0 Y Cathy 5200 Y

brulard · ‎03-20-2017

hello, I need to transform data for the purpose of evaluating the actions that employees have taken on customers' account. It would be greatly appreciated if someone could point me in the right direction. I'm looking at previous month's data, which captures on a given day, any time an agent "actions" / modifies a customer's account. The action taken on customer's account are categorized by risk level, from 1 (most critical) to 3 less, and 0, no risk. What I need to do: Flag 5 observations per agent of which 3 have a risk_level=1, and 2 have a risk_level=2. If cannot find total of 5 obs, then select from risk_level=3, and if still can't find, select risk_level=0. (The goal is to flag 5 where 3 have a risk level of 1, and two a lower risk level) Ultimately, it would be great if I could export the data to excel, and have a pivot table for agent_name, and variable Subject_to_review, where each agent name would have a count 5. Example data has been added below. Thanks

brulard · ‎03-16-2017

hi, Does someone have suggestion code to offer on how I can assign a dummy variable (1 or 0) to identify a random sample from a population? Say I have a population of 1,800 records and I want to flag exactly 35, randomly. The random sample volume will always remain the same. I want to retain the full 1,800 records, but with a column that can be filtered to, say, 1, to identify the random sample. Thank you

brulard · ‎03-13-2017

ok, I think i figured it out. The result I was getting, not true false postives... but is the result of my string |PIN| which flagged word SHOPPING. To avoid hitting this, i could add a space before and after, | PIN |. So i think this closes this message

brulard · ‎03-13-2017

thanks... issue is primarily finding whole word match... not sure why I am getting false positves

brulard · ‎03-10-2017

thanks for the tip... i ll try it on Monday

brulard · ‎03-10-2017

Hi Ballardw, here is an example: MDECLINED LOC INCR TO 4K$ FOR X-MAS SHOPPING, SCORE 684/3/471, EST ON BURO SINCE 1990, NO DEROGS ON BURO, DEBT RATIO HIGH, HIGHEST TRADE 18K$ ALL MAX ON OTHER TRADE, GOOD PYMT HISTORY.

brulard · ‎03-10-2017

hi I'm trying to scan a column (using base SAS) that contains notations that agents enter when dealing with customers. The problem is that i am getting many false positives. That is, when coding a string of words to flag, I get records that flag 1, yet without the words being in the value. I tried using both index function and the prxmatch function, with same result. The column that I am scanning is formatted as $2000. Example of my query: data want;set have: if prxmatch("m/COUNTER|COUNTERS|CW30|REISSUE|REISSUED|STRATEGY|RE-OPEN|RE-ISSUE|APPROVE|PIN|SECURITY|TRANSACTION|TRANSACTIONS/i",column_have)> 0 then found=1; else found=0;RUN; If you have a suggestion as an alternate way to flag for presence of words in given field, with better accuracy, please advise. thank you

brulard · ‎03-08-2017

that works too. Thanks Ksharp

brulard · ‎03-07-2017

this is very helpful, thanks again

brulard · ‎03-07-2017

hi, I need to count times string "RE" appears in multicharacter string. String is always char. Example: Data_Want Column_have_id Column_have_string 2 01 NDCSPCRCBCSLAFDWIWOFDSPDERRDETDRECADETDRECA 1 02 ENDCSPCRCBCSLAFDWIWOFDSPDERRDETDRECADETD 1 03 ENDCSPCRCBCSLAFDWIWOFDSPDERRDETDRECADETD 0 04 ENDCSPCRCBCSLAFDWIWOFDSP In the first string of data in the above example, "RE" is read twice. ("RE" should be captured as is, not "ER"...). thank you

brulard · ‎03-03-2017

hi nehalsanghvi, if there are multiple rows for any ID, but not all rows have a value in Var1 (some do), then what would you want the new_dummy_var to be? The new_dummy_var to be set to 1 only if there are no values in Var1 at all. Thanks,

brulard · ‎03-03-2017

It could be a number or char. thanks

brulard · ‎03-02-2017

hi, can someone recommend some code so that sas check for condition: if value IDs is duplicate [could be more than 1 duplicate], then scan Var1 for content, if no content, create new_dummy_var=1. If value IDs not duplicate, skip. Var1 is numeric. UPDATING Example outcome: IDs Var 1 new_dummy_var A001 5 A001 6 A001 5 A002 1 A002 1 A003 5 A003 A004 **Update: In the example outcome, note that because one of the duplicate observation A003 has a 5 in Var1, the program should move on, and not assign a value in New_dummy_var. **New_dummy_var is only meant to be a flag, so not care if char or numeric. Thank you very much Dave

Online Status	Offline
Date Last Visited	‎03-01-2021 10:19 PM

Re: Vertical align text and graphic in banner

Re: Vertical align text and graphic in banner

Vertical align text and graphic in banner

Re: Macro to export each table in a library to an Excel file

Re: Macro to export each table in a library to an Excel file

Macro to export each table in a library to an Excel file

Re: guidance to produce donut with percent outer, count inner

guidance to produce donut with percent outer, count inner

Re: Concat by maintaining space or adding between strings

Re: Concat by maintaining space or adding between strings

Re: Excluding rows when two variables have specific values on the same...

Re: ODS, EXCELXP, PROC REPORT, VJUST COLUMN

Re: how to pad character variable with leading zeroes?

Re: Get a list of dataset names in a directory/library

Re: Vertical align text and graphic in banner

Re: Error running script in Hadoop (Apache)

best practice for daily query fetch date

Re: Hadoop connectivity issue

Re: Flag random data based on criteria

Re: Flag random data based on criteria

Re: Flag random data based on criteria

Flag random data based on criteria

Add dummy column to flag random sample from population

Re: Getting false positives when flagging for words

Re: Getting false positives when flagging for words

Re: Getting false positives when flagging for words

Re: Getting false positives when flagging for words

Getting false positives when flagging for words

Re: Count presence of two-character string among many characters

Re: Count presence of two-character string among many characters

Count presence of two-character string among many characters

Re: if duplicate, check if var1 has content, if no then flag

Re: if duplicate, check if var1 has content, if no then flag

if duplicate, check if var1 has content, if no then flag