Converting R Code to SAS

[asherer]

Wondering if anyone has a good site to convert R code to SAS code. I was wondering if I can get help converting the following code to SAS. Any guidance is appreciative. require(dplyr) source("bisg_proxy/new_func.R") ia_s <-read.csv('ready_for_r.csv', colClasses = 'character') #add geo tract and geo block ia_2<- ia_s %>% mutate(geo_tract=paste(State, County, Censustrac, sep=''), geo_tract=gsub('[.]','',geo_tract))%>% mutate(geo_block=paste(State, County, Censustrac, BlockGrp, sep=''), geo_block=gsub('[.]','',geo_block)) # if blockgrp == 0, BISG at the tract level #note duplicated primary as secondary for name/dob as no secondary in file # if blockgrp == 0, BISG at the tract level ia_t <- ia_2 %>% filter(BlockGrp %in% c("0",NA)) %>% applyProxy(pri_first_name = 'pri_first_name', pri_last_name = 'pri_last_name', sec_first_name = 'sec_first_name', geoid = 'geo_tract', geoid_key = 'GeoInd', threshold = .75, keepall = T, census_geog = tract_attr_over18) # if blockgrp != 0, BISG at the block group level ia_b <- ia_2 %>% filter(BlockGrp != '0') %>% applyProxy(pri_first_name = 'pri_first_name', pri_last_name = 'pri_last_name', sec_first_name = 'sec_first_name', geoid = 'geo_block', geoid_key = 'geoind', threshold = .75, keepall = T, census_geog = blkgrp_attr_over18) ia_p <- rbind(ia_t, ia_b) %>% arrange(race, gender) #update header names if applicable ia_p_s <- ia_p %>% select(application_id, gender, race, pri_femaleprob:sec_maleprob, Probability_Asian:Probability_White, pri_address, pri_city, pri_state, pri_zip, State, County, Censustrac, BlockGrp, MSA, pri_first_name, pri_last_name) # this exports your file into your folder as a .csv write.csv(ia_p_s, file='proxied.csv', row.names=FALSE

[PaigeMiller]

You need to re-format your message so it is readable.

Click on the {i} icon and paste your R-code into that window. Make sure you have one command on one line; not many commands on one line.

[Reeza]

There is no automatic conversion as far as I know. Can you post your code reformatted to be legible and then maybe we can offer some options? 

I think if you post your code using a code block (6/7th icon) in the Rich Text editor. 

---

@asherer wrote:

Wondering if anyone has a good site to convert R code to SAS code. I was wondering if I can get help converting the following code to SAS. Any guidance is appreciative. require(dplyr) source("bisg_proxy/new_func.R") ia_s <-read.csv('ready_for_r.csv', colClasses = 'character') #add geo tract and geo block ia_2<- ia_s %>% mutate(geo_tract=paste(State, County, Censustrac, sep=''), geo_tract=gsub('[.]','',geo_tract))%>% mutate(geo_block=paste(State, County, Censustrac, BlockGrp, sep=''), geo_block=gsub('[.]','',geo_block)) # if blockgrp == 0, BISG at the tract level #note duplicated primary as secondary for name/dob as no secondary in file # if blockgrp == 0, BISG at the tract level ia_t <- ia_2 %>% filter(BlockGrp %in% c("0",NA)) %>% applyProxy(pri_first_name = 'pri_first_name', pri_last_name = 'pri_last_name', sec_first_name = 'sec_first_name', geoid = 'geo_tract', geoid_key = 'GeoInd', threshold = .75, keepall = T, census_geog = tract_attr_over18) # if blockgrp != 0, BISG at the block group level ia_b <- ia_2 %>% filter(BlockGrp != '0') %>% applyProxy(pri_first_name = 'pri_first_name', pri_last_name = 'pri_last_name', sec_first_name = 'sec_first_name', geoid = 'geo_block', geoid_key = 'geoind', threshold = .75, keepall = T, census_geog = blkgrp_attr_over18) ia_p <- rbind(ia_t, ia_b) %>% arrange(race, gender) #update header names if applicable ia_p_s <- ia_p %>% select(application_id, gender, race, pri_femaleprob:sec_maleprob, Probability_Asian:Probability_White, pri_address, pri_city, pri_state, pri_zip, State, County, Censustrac, BlockGrp, MSA, pri_first_name, pri_last_name) # this exports your file into your folder as a .csv write.csv(ia_p_s, file='proxied.csv', row.names=FALSE

---

[asherer]

Sorry not sure what I did when I pasted it the first time.

require(dplyr)

source("bisg_proxy/new_func.R")

ia_s <-read.csv('ready_for_r.csv', colClasses = 'character')

#add geo tract and geo block

ia_2<- ia_s %>%
  mutate(geo_tract=paste(State, County, Censustrac, sep=''),
         geo_tract=gsub('[.]','',geo_tract))%>%
  mutate(geo_block=paste(State, County, Censustrac, BlockGrp, sep=''),
         geo_block=gsub('[.]','',geo_block))

# if blockgrp == 0, BISG at the tract level
#note duplicated primary as secondary for name/dob as no secondary in file

# if blockgrp == 0, BISG at the tract level

ia_t <- ia_2 %>%
  filter(BlockGrp %in% c("0",NA)) %>%
  applyProxy(pri_first_name = 'pri_first_name',
             pri_last_name = 'pri_last_name',
             sec_first_name = 'sec_first_name',
             geoid = 'geo_tract',
             geoid_key = 'GeoInd',
             threshold = .75,
             keepall = T,
             census_geog = tract_attr_over18)

# if blockgrp != 0, BISG at the block group level
ia_b <- ia_2 %>%
  filter(BlockGrp != '0') %>%
  applyProxy(pri_first_name = 'pri_first_name',
             pri_last_name = 'pri_last_name',
             sec_first_name = 'sec_first_name',
             geoid = 'geo_block',
             geoid_key = 'geoind',
             threshold = .75,
             keepall = T,
             census_geog = blkgrp_attr_over18)

ia_p <- rbind(ia_t, ia_b) %>%
  arrange(race, gender)

#update header names if applicable
ia_p_s <- ia_p %>%
  select(application_id, gender, race, pri_femaleprob:sec_maleprob,
         Probability_Asian:Probability_White, pri_address, pri_city, pri_state, pri_zip, State, County, Censustrac, BlockGrp, MSA, pri_first_name, pri_last_name)

# this exports your file into your folder as a .csv
write.csv(ia_p_s, file='proxied.csv', row.names=FALSE)

[PaigeMiller]

The preferred form is as I described earlier, "Click on the {i} icon and paste your R-code into that window."

Could you please do that, this makes code much easier to read.

[asherer]

require(dplyr)
#require(openxlsx)
source("bisg_proxy/new_func.R")


ia_s <-read.csv('Overrides for R.csv', colClasses = 'character')


#ia_s <- read.xlsx('Policy Exceptions for R.xlsx')

#add geo tract and geo block

ia_2<- ia_s %>%
  mutate(geo_tract=paste(state, county, censustrac, sep=''),
         geo_tract=gsub('[.]','',geo_tract))%>%
  mutate(geo_block=paste(state, county, censustrac, blockgrp, sep=''),
         geo_block=gsub('[.]','',geo_block))



# if blockgrp == 0, BISG at the tract level
#note duplicated primary as secondary for name/dob as no secondary in file

# if blockgrp == 0, BISG at the tract level

ia_t <- ia_2 %>%
  filter(blockgrp %in% c("0",NA)) %>%
  applyProxy(pri_first_name = 'pri_first_name', 
             pri_last_name = 'pri_last_name',
             sec_first_name = 'sec_first_name',
             geoid = 'geo_tract',
             geoid_key = 'GeoInd',
             threshold = .5,
             keepall = T,
             census_geog = tract_attr_over18)


# if blockgrp != 0, BISG at the block group level
ia_b <- ia_2 %>%
  filter(BlockGrop != '0') %>%
  applyProxy(pri_first_name = 'pri_first_name', 
             pri_last_name = 'pri_last_name',
             sec_first_name = 'sec_first_name',
             geoid = 'geo_block',
             geoid_key = 'geoind',
             threshold = .5,
             keepall = T,
             census_geog = blkgrp_attr_over18)

ia_p <- rbind(ia_t, ia_b) %>%
  arrange(race, gender)

#update header names if applicable 
ia_p_s <- ia_p %>%
  select(application_id, gender, race, pri_femaleprob:sec_maleprob, 
         Probability_Asian:Probability_White, pri_address, pri_city, pri_state, pri_zip, state, county, censustrac, blockgrp, msa, pri_first_name, pri_last_name)

# this exports your file into your folder as a .csv
write.csv(ia_p_s, file='Policy Exceptionsw no DOB proxied.csv', row.names=FALSE)

[Reeza · Accepted solution]

I'm familiar with all except applyProxy, do you have a reference for that command, not finding it online either. From the package name it appears to be a fuzzy match for names resulting in probability of gender/ethnicity? I don't think SAS will have a replacement for that.

Either way, most appear to be select or computations, so you can easily use Base SAS to do all of this.

Video.sas.com has most of what you need.

PROC IMPORT will read CSV or Excel files.

CATX() function replaces paste()
WHERE will filter and can be used in almost all procedures.
SAS has had IN for a long time so that's comparable.

[Reeza]

SAS DataFlux may be able to do some of this. And of course the accuracy depends highly on the data and where you're using it. Due to the racial demographics of where I live, it would be almost useless.