SWMM .inp File loses information by writing, running and reading in a for loop

[fcfunke]

Hi everyone!
I am struggling with a problem reading, writing and running the same SWMM .inp File in a for-loop several times.
The problem appears in the $outfalls part of the SWMM .inp File, affecting the "Tide Gate" and "Route To" parameters.
The data is switching columns in the tibble and losing information, resulting in a SWMM error.
I reproduced the problem by reading and writing the same .inp file 3 times without changing anything. (see below)
Interestingly if I run the .inp file without the "Route To" Parameter (SG31) it still works.

`> input_test <- swmmr::read_inp(x = "D://Uni_Innsbruck/Dauerstufen_Jaehrlichkeiten_Analyse/SWMM_Modell_Berlin/inp_test/Mulde5_0.inp")
Warnmeldungen:
1: Expected 6 pieces. Additional pieces discarded in 14 rows [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].
2: Expected 9 pieces. Additional pieces discarded in 14 rows [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14].

input_test$outfalls

A tibble: 3 x 6

Name Elevation Type Stage Data Gated Route To

1 Out5 46.7 FREE NA O ""
2 Out2 47.6 FREE NA NO "SG31"
3 Out1 0 FREE NA NO ""

swmmr::write_inp(input_test, "D://Uni_Innsbruck/Dauerstufen_Jaehrlichkeiten_Analyse/SWMM_Modell_Berlin/inp_test/Mulde5_0.inp")
input_test <- swmmr::read_inp(x = "D://Uni_Innsbruck/Dauerstufen_Jaehrlichkeiten_Analyse/SWMM_Modell_Berlin/inp_test/Mulde5_0.inp")
input_test$outfalls

A tibble: 3 x 6

Name Elevation Type Stage Data Gated Route To

1 Out5 46.7 FREE "O" NA NA NA
2 Out2 47.6 FREE "NO SG31" NA NA NA
3 Out1 0 FREE NO "" NA NA NA

swmmr::write_inp(input_test, "D://Uni_Innsbruck/Dauerstufen_Jaehrlichkeiten_Analyse/SWMM_Modell_Berlin/inp_test/Mulde5_0.inp")
input_test <- swmmr::read_inp(x = "D://Uni_Innsbruck/Dauerstufen_Jaehrlichkeiten_Analyse/SWMM_Modell_Berlin/inp_test/Mulde5_0.inp")
input_test$outfalls

A tibble: 3 x 6

Name Elevation Type Stage Data Gated Route To

1 Out5 46.7 FREE "" NA NA NA
2 Out2 47.6 FREE "O SG31" NA NA NA
3 Out1 0 FREE NO "" NA NA NA `

Thanks for any help!

with best regards
Fabian

[dleutnant]

@fcfunke Thanks for spotting this. I somehow can reproduce the issue with the SWMM examples shipped with the swmmr package. However, I am currently on paternity leave and thus it may take a while.

[dleutnant]

issue potentially here:

swmmr/R/parse_section.R

Lines 224 to 233 in 3e373e6

	parse_section.outfalls <- function(x, ...) {
	separate_into(x, c("Name", "value")) %>%
	separate_into(
	sep = base::cumsum(c(10, 1, 10, 1, 10, 1, 10, 1)), extra = "warn",
	fill = "warn", into = c(
	"Elevation", "tab2", "Type", "tab3", "Stage Data", "tab4", "Gated",
	"tab5", "Route To"
	))
	}

[dleutnant]

@fcfunke:
Could you please substitute NA with ' ' (double whitespace character) before you write the .inp file? And then read again?
The issue might be related to the write_inp function and the way NA values are handled.

[fcfunke]

@dleutnant
Sorry to come back to the problem so late, helped myself by saving the input$outfalls tibble before the first run and then just changing it every run to the initial version.
Today I stumbled over the problem again and made the following discovery. Not sure if this helps to solve the problem.

When reading the input file for the first time, input$outfalls$Elevation is of type "dbl" and just contains the Elevation values:

input$outfalls

A tibble: 20 x 6

Name Elevation Type Stage Data Gated Route To

1 2DOut 373. FREE NA NO NA
2 OF1 377. FREE NA NO NA
3 OF10 377. FREE NA NO NA

After the first write, run and read process input$outfalls$Elevation is of type "chr" and contains both Elevation and Type values:

input$outfalls

A tibble: 20 x 6

Name Elevation Type Stage Data Gated Route To

1 2DOut 373.1 FREE NO NA NA NA
2 OF1 376.6 FREE NO NA NA NA
3 OF10 376.7 FREE NO NA NA NA

with best regards
Fabian

[dleutnant]

Thanks for your report. Would you be able to file a reproducible example using one of SWMM's model examples shipped with the package?