get_SDA_interpretation sometimes throws warning for NCCPI

[kevinwolz]

I'm using the approach developed in #303 to get NCCPI data. get_SDA_interpretation sometimes throws the below warning for me and sometimes not, depending on the specific mukeys passed to it. Attached here is an Rdata file (in the zip file) containing a vector of mukeys that does generate the warning. What's going on here?

mukeys <- readRDS("mukeys.RData")
out <- get_SDA_interpretation(rulename    = "NCCPI - National Commodity Crop Productivity Index (Ver 3.0)",
method      = "Dominant Component",
mukeys      = mukeys,
wide_reason = TRUE) 

Warning message:
In rbind(c("NCCPI - NCCPI Corn Submodel (I)", "Not rated"), c("NCCPI - NCCPI Small Grains Submodel (II)",  :
  number of columns of result is not a multiple of vector length (arg 1)

mukeys.RData.zip

[brownag]

Hey @kevinwolz

This appears to be an unhandled issue with some components that lack ratings for most of the subrules.

I am unable to open the attached .RData file. But I can reproduce the issue here with mukey 459419 which is CA041 mapunit symbol 127 "Fluvents, channeled".

It is 100% higher taxa "Fluvents" and is "Not rated" for anything except "Impacted soil" subrule. The issue above arises because the impacted soil has a numeric rating of zero, whereas the others have "Not rated" reason with no numeric, which creates a non-conformal rbind condition that kinda messes things up with the resulting character matrix. Thanks for pointing this out

library(soilDB)
y <- get_SDA_interpretation(
  rulename    = "NCCPI - National Commodity Crop Productivity Index (Ver 3.0)",
  method      = "Dominant Component",
  mukeys      = 459419,
  wide_reason = TRUE
)
#> Warning in rbind(c("NCCPI - NCCPI Corn Submodel (I)", "Not rated"), c("NCCPI -
#> NCCPI Small Grains Submodel (II)", : number of columns of result is not a
#> multiple of vector length (arg 1)
y
#>    mukey    cokey areasymbol musym              muname compname
#> 1 459419 24586388      CA041   127 Fluvents, channeled Fluvents
#>             compkind comppct_r majcompflag
#> 1 Taxon above family       100         Yes
#>   rating_NCCPINationalCommodityCropProductivityIndexVer30
#> 1                                                      NA
#>   class_NCCPINationalCommodityCropProductivityIndexVer30
#> 1                                              Not rated
#>                                                                                                                                                                                             reason_NCCPINationalCommodityCropProductivityIndexVer30
#> 1 NCCPI - NCCPI Corn Submodel (I) "Not rated" (); NCCPI - NCCPI Small Grains Submodel (II) "Not rated" (); NCCPI - NCCPI Cotton Submodel (II) "Not rated" (); NCCPI - NCCPI Soybeans Submodel (I) "Not rated" (); Impacted soil "No limitation" (0)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPICornSubmodelI
#> 1                                                        NCCPI - NCCPI Corn Submodel (I)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPISmallGrainsSubmodelII
#> 1                                                       NCCPI - NCCPI Small Grains Submodel (II)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPICottonSubmodelII
#> 1                                                        NCCPI - NCCPI Cotton Submodel (II)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPISoybeansSubmodelI
#> 1                                                        NCCPI - NCCPI Soybeans Submodel (I)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_Impactedsoil
#> 1                                                                           0

After 1ac18dc I get no warnings but let me know if you find others that have this kind of issue.

library(soilDB)
y <- get_SDA_interpretation(
  rulename    = "NCCPI - National Commodity Crop Productivity Index (Ver 3.0)",
  method      = "Dominant Component",
  mukeys      = 459419,
  wide_reason = TRUE
)
y
#>    mukey    cokey areasymbol musym              muname compname
#> 1 459419 24586388      CA041   127 Fluvents, channeled Fluvents
#>             compkind comppct_r majcompflag
#> 1 Taxon above family       100         Yes
#>   rating_NCCPINationalCommodityCropProductivityIndexVer30
#> 1                                                      NA
#>   class_NCCPINationalCommodityCropProductivityIndexVer30
#> 1                                              Not rated
#>                                                                                                                                                                                             reason_NCCPINationalCommodityCropProductivityIndexVer30
#> 1 NCCPI - NCCPI Corn Submodel (I) "Not rated" (); NCCPI - NCCPI Small Grains Submodel (II) "Not rated" (); NCCPI - NCCPI Cotton Submodel (II) "Not rated" (); NCCPI - NCCPI Soybeans Submodel (I) "Not rated" (); Impacted soil "No limitation" (0)
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPICornSubmodelI
#> 1                                                                                   <NA>
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPISmallGrainsSubmodelII
#> 1                                                                                           <NA>
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPICottonSubmodelII
#> 1                                                                                      <NA>
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_NCCPINCCPISoybeansSubmodelI
#> 1                                                                                       <NA>
#>   rating_reason_NCCPINationalCommodityCropProductivityIndexVer30_Impactedsoil
#> 1                                                                           0

[kevinwolz]

Thanks for the quick response! The mukeys I was working with no work on my end without a warning.

[brownag]

Great. Closing this, but we can continue fixing any issues with wide_reason subroutine on #316 where you uncovered a much more general problem.