apply on dataframe works for one row, but apply over multiple rows fails for that particular row

Apply on dataframe works for one row, but apply over multiple rows fails for that particular row Here's an example (row 954 works in isolation):

     apply(dx[954,], 1, query.db)
            2571
1092  0.06044147
3890  0.05948577
3273  0.05911261
794   0.05855583
4241  0.05854064
3560  0.05602690

But when it's combined with one or more of the rest of the rows in the data.frame, row 954 returns numeric(0):

 apply(dx[954:955,], 1, query.db)
$`2571`
numeric(0)

$`1194`
         908         3754         5119         2835         3139         2775
 0.028935504  0.026941886  0.024762815  0.023335049  0.021892711  0.021001654
        1101         1762          401         3899         3727         4248
 0.020434257  0.020413724 -0.083333333  0.020111260  0.019872379  0.019840807
        3282         4585          798         5249          836         3570
 0.019729315  0.019524213  0.019440097  0.019410513  0.019288864  0.019052412
         726         3663         4775         3030         4938         3259
 0.018894556  0.018838258  0.018749506 -0.083333333  0.018416353  0.018240173
        1426         4563         3871          958         1242         1033

Here are the functions I'm using (please forgive the <<-, needed access to those variables in separate functions! Any suggestions on re-writing that are welcome! You can see my attempt at re-writing an an-line function which did not work and is commented out.

query.db <- function(dxdf){
    snquery = dbSendQuery(connection, paste("select distinct cpt4code from DxRxLookup where icd9code = '", dxdf[4], "'", sep = ""))
    cptcodes <<- dbGetResult(snquery)
    if(dim(cptcodes)[1] == 0){
        return (0)
    }

    #restrict this to constrain by Rx1 and Rx2 inputs
    h1 <- which(cpt$Cpt %in% cptcodes[,1])
    y <- which((cpt$Rx2 %in% dxdf[2] | cpt$Rx1 %in% dxdf[2]) & (cpt$Rx1 %in% dxdf[3] | cpt$Rx2 %in% dxdf[3]))
    hits <- intersect(y, h1)

    sum.all.hits <<- sum(cpt[hits,开发者_如何学C]$Log.Odds.Ratio)
    miss <<- 1 / (length(y) - length(hits))
#   apply(cpt[y,], 1, function(cptdf) if (cptdf[4] %in% codes[,1]){ P.hit <- as.numeric(cptdf[6]) / sum.of.hits; return (P.hit)} else { return (-m)}, c(codes=cptcodes, sum.of.hits=sum.all.hits, m=miss))
    apply(cpt[y,], 1, calc.enrichment)

}

Are you sure your SQL query returns what you would like? Numeric(0) suggest that you retrieve an empty resultset. Furthermore, as pointed by @Prasad Chalasani, try to post a more precise problem description. From the above I am not sure I understand what your aim is.