CRHM GroupEnding name in AZ, BZ, CZ, DZ etc

[loganxingfang]

I have a CRHM user that is running a CRHM group prj. The group prj has 150 groups, ie. from Grp@A to GrpET, and when the group prj finished simulation and saved output, the variables in Sub052_Grp@AZ , Sub078_Grp@BZ, Sub104_Grp@CZ, Sub130_Grp@DZ cannot be saved in the output. I think there is some code missing in crhmcode that handles the large number of groups, ie. missing code to build and read these groups. I attached the group prj along with obs files used here
obs_files_CMFD_20072008.zip

Borland CRHM had this issue before it was fixed in February 2019. I am not the original programmer that fixed this, but I believe there are two files in borland_source folder:

CRHMmain.cpp (https://github.com/srlabUsask/crhmcode/blob/master/borland_source/CRHMmain.cpp)

Common.cpp (https://github.com/srlabUsask/crhmcode/blob/master/borland_source/Common.cpp)

In both files, there are codes for GroupEnding. I am not sure how these are coded in crhmcode.

[jhs507]

Looks like this is as you suspected a case of code not getting updated between versions.

This is the code for group endings in borland

void GroupEnding(string &AA, int Cnt){
  int alphas = Cnt/26;
  int letter = Cnt%26;
  if(!letter){ // 26th letter is 'Z' not "@@"
    letter = 26;
    alphas -= 1;
  }
  AA = "@@";
  AA[1] += letter;
  if(alphas > 0){
    string B = "@";
    B[0] += alphas;
    AA.insert(1, B);
  }
}

This is the code in crhmcode:

void Common::GroupEnding(string &AA, int Cnt) {

	int alphas = Cnt / 26;
	int letter = Cnt % 26;
	if (!letter) { // 26th letter is 'Z' not "@@"
		letter = 26;
		alphas = 0;
	}
	AA = "@@";
	AA[1] += letter;
	if (alphas > 0) {
		string B = "@";
		B[0] += alphas;
		AA.insert(1, B);
	}
}

Looks like an easy fix to just add the code that allows for AZ BZ ect to be generated as group endings. I will implement it and send you a version of CRHM with the fix for you to verify it.

[loganxingfang]

Looks like this is as you suspected a case of code not getting updated between versions.

This is the code for group endings in borland

void GroupEnding(string &AA, int Cnt){
  int alphas = Cnt/26;
  int letter = Cnt%26;
  if(!letter){ // 26th letter is 'Z' not "@@"
    letter = 26;
    alphas -= 1;
  }
  AA = "@@";
  AA[1] += letter;
  if(alphas > 0){
    string B = "@";
    B[0] += alphas;
    AA.insert(1, B);
  }
}

This is the code in crhmcode:

void Common::GroupEnding(string &AA, int Cnt) {

	int alphas = Cnt / 26;
	int letter = Cnt % 26;
	if (!letter) { // 26th letter is 'Z' not "@@"
		letter = 26;
		alphas = 0;
	}
	AA = "@@";
	AA[1] += letter;
	if (alphas > 0) {
		string B = "@";
		B[0] += alphas;
		AA.insert(1, B);
	}
}

Looks like an easy fix to just add the code that allows for AZ BZ ect to be generated as group endings. I will implement it and send you a version of CRHM with the fix for you to verify it.

In the Borland Common.cpp file, there are two blocks of code for GroupEndings

Block 1 from line 20 to 34,
void GroupEnding(string &AA, int Cnt){
int alphas = Cnt/26;
int letter = Cnt%26;
if(!letter){ // 26th letter is 'Z' not "@@"
letter = 26;
alphas -= 1;
}
AA = "@@";
AA[1] += letter;
if(alphas > 0){
string B = "@";
B[0] += alphas;
AA.insert(1, B);
}
}

Blocks 2 from line 36 to 52,
void GroupEnding(String &AA, int Cnt){
int alphas = Cnt/26;
int letter = Cnt%26;

if(!letter){ // 26th letter is 'Z' not "@@"
letter = 26;
alphas -= 1;
}

AA = "@@";
AA[2] += letter;
if(alphas > 0){
String B = "@";
B[1] += alphas;
AA.Insert(B, 2);
}
}

Would the block 2 be also needed in crhmcode Common.cpp file?

[jhs507]

Yes I noticed that and didn't know what to make of it. They have the exact same signatures so it doesn't make sense for both of them to be present. Looking them over I suspect that the first one is the correct definition and I can't imagine that borland crhm is using both somehow.

[loganxingfang]

Yes I noticed that and didn't know what to make of it. They have the exact same signatures so it doesn't make sense for both of them to be present. Looking them over I suspect that the first one is the correct definition and I can't imagine that borland crhm is using both somehow.

What I noticed the differences are:

In the first blocks, these are used:

AA[1] += letter;

B[0] += alphas;

AA.insert(1, B);

While in the second block, these are used:

AA[2] += letter;

B[1] += alphas;

AA.Insert(B, 2);

[jhs507]

Yes, the second code doesn't make sense to me as it adds the group letters after @@ instead of after a single @

[jhs507]

In general I don't like how this code handles adding to strings as it makes the code quite hard to read.

[jhs507]

Let me know how the test goes with the version I sent you.

[loganxingfang]

Let me know how the test goes with the version I sent you.

Thanks Justin. I will pass on that installation file to the CRHM user and will let you know how this version does.

[loganxingfang]

Let me know how the test goes with the version I sent you.

@jhs507 I got feedback and see the following:

"I tested the new version of CRHM on group running. It failed, there is no error message. The software just closed by itself without any output.

I think it may be caused by the different versions. But I reset the group running from the very beginning. The same problem appears.

It could run at one single subbasin.

The new version has no End Date in the platform. It may also have limited output variables. When I set the group using Macro, it seems that only I can select the subbasin, but could not enter any words in the Macro."

[jhs507]

I am a little confused by this feedback. If the researcher could reply directly here that may be helpful.

I attempted to run it myself and found that because it was trying to add over 1000 series to the TChart Display this caused an error. There is feedback in CRHMcode GUI that advises the user that when they have selected over 200 inputs that it can become unstable. I think this is a case of that causing issues. I would advise attempting to run this in CRHMcode gcc or without using the GUI to see if it can produce expected results as CRHMcode GUI is not intended to be able to display so many outputs on the graph.

I think it may be caused by the different versions. But I reset the group running from the very beginning. The same problem appears.

I don't understand this comment what was reset?

The new version has no End Date in the platform. It may also have limited output variables. When I set the group using Macro, it seems that only I can select the subbasin, but could not enter any words in the Macro

I also don't understand this comment, what is meant by no End Date in the platform and what is meant by can not enter words in the Macro.

Would it be possible to test a more simple model that still has groups up to AZ, BZ, CZ ect? as I would like to confirm that the group ending fix I implemented worked before trying to address other possible issues.