cantera/tools/testtools/csvdiff.cpp

/*
 *  csvdiff File1.csv File2.csv
 *
 *  Compares the variable values in two Excel formatted
 *  comma separated files.
 *  The comparison is done using a weighted norm basis.
 *
 *  The two files should be basically equal. However, File1.csv is
 *  taken as the reference file, that has precedence, when there is
 *  something to be decided upon.
 *
 *  Arguments:
 *   -h = prints this usage information
 *
 *  Shell Return Values
 *    0 = Comparison was successful
 *    1 = One or more nodal values failed the comparison
 *    2 = One or more of the header values failed the comparison
 *    3 = Apples to oranges, the files can not even be compared against
 *        one another.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include <float.h>
#include <limits.h>
#include "kernel/config.h"
#ifndef _MSC_VER
#include <unistd.h>
#else
#include <string>
#endif
using namespace std;



#if defined(__CYGWIN__)
#include <getopt.h>
#endif

#include "mdp_allo.h"
#include "tok_input_util.h"

#ifndef MAX
#  define MAX(x,y)    (( (x) > (y) ) ? (x) : (y))
#endif
#ifndef MIN
#  define MIN(x,y)    (( (x) < (y) ) ? (x) : (y))
#endif

int Debug_Flag = true;
double grtol = 1.0E-3;
double gatol = 1.0E-9;

#define RT_PASSED  0
#define RT_FAILED_COL 1
#define RT_FAILED_HDR 2
#define RT_FAILED_OTHER 3

/*
 * First iteration towards getting this variable
 */
int Max_Input_Str_Ln = MAX_INPUT_STR_LN;
/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

#ifdef _MSC_VER
/*
 * Windows doesn't have getopt(). This is an incomplete version that
 * does enough to handle required functionality.
 */
int optind = -1;
char *optarg = 0;

int getopt(int argc, char **argv, const char *) {
    static int currArg = 1;
    static int currOptInd = 1;
    string tok;
    static int charPos = 0;
    int rc = -1;
    if (currArg >= argc) {
      optarg = 0;
      return -rc;
    }
    tok = string(argv[currArg]);
    currOptInd = currArg+1;
    if (currOptInd > argc - 1) {
      currOptInd = -1;
      optarg = 0;
    } else {
      optarg = argv[currArg+1];
    }
    size_t len = strlen(tok.c_str());
    if (charPos == 0) {
      bool found = false;
      do {
	tok = string(argv[currArg]);
	len = strlen(tok.c_str());
	if (len > 1 && tok[0] == '-') {
	  found = true;
	  charPos = 1;
	  if (len > 2 && tok[1] == '-') {
	    charPos = 2;
	  }
	} else {
	  if (optind == -1) {
	    optind = currArg;
	  }
	}
	if (!found) {
	  if (currArg < (argc-1)) {
	    currArg++;
	  } else {
            optarg = 0;
	    return -1;
	  }
	}
      } while (!found);
    }
 
    rc = tok[charPos];
    if (charPos < static_cast<int>(len - 1)) {
      charPos++;
    } else {
      charPos = 0;
    }
    return rc;
}

#endif

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static int diff_double(double d1, double d2, double rtol, double atol)

/*
 * Compares 2 doubles. If they are not within tolerance, then this
 * function returns true. 
 */
{
  if (fabs(d1-d2) > (atol + rtol * 0.5 * (fabs(d1) + fabs(d2)))) return 1;
  return 0;
}

static int diff_double_slope(double d1, double d2, double rtol, 
    double atol, double xtol, double slope1, double slope2)

/*
 * Compares 2 doubles. If they are not within tolerance, then this
 * function returns true. 
 */
{
    double atol2 = xtol*(fabs(slope1) + fabs(slope2));
  if (fabs(d1-d2) > (atol + atol2 + rtol * 0.5 * (fabs(d1) + fabs(d2)))) return 1;
  return 0;
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static double calc_rdiff(double d1, double d2, double rtol, double atol)

/*
 * Calculates the relative difference using a fuzzy comparison
 */

{
  double rhs, lhs;
  rhs = fabs(d1-d2);
  lhs = atol + rtol * 0.5 * (fabs(d1) + fabs(d2));
  return (rhs/lhs);
}

/*****************************************************************************/
/*
 * breakStrCommas(): 
 *    This routine will break a character string into stringlets according
 *    to the placement of commas. The commas are replaced by null 
 *    characters. 
 *
 *    Argument:
 *      str => original string. On exit, this string will have beent
 *             altered.
 *      strlets -> Vector of pointers to char *.  The vector has a size
 *                 larger than or equal to maxPieces.
 *      maxPieces -> largest number of pieces to divide the string into.
 *
 *    Return:
 *      This returns the number of pieces that the string is actually
 *      broken up into.
 */

static int breakStrCommas(char *str, char **strlets, int maxPieces)
{
  int numbreaks = 0;
  if (strlets) {
    strlets[0] = str;
    if (str) {
      char *cptr = str;
      char *cetn = NULL;
      do {
	cetn = strchr(cptr, (int) ',');
	if (cetn) {
	  numbreaks++;
	  cptr = cetn + 1;
	  strlets[numbreaks] = cptr;
	  *cetn = '\0';
	}
      } while (cetn && (numbreaks < (maxPieces - 1)));
    }
  }
  return numbreaks + 1;
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/
#define LT_NULLLINE  0
#define LT_TITLELINE 1
#define LT_COLTITLE  2
#define LT_DATALINE  3
/*
 * get_sizes()
 * 
 * This routine obtains the sizes of the various elements of the file
 * by parsing the file.
 *   (HKM: Note, this file could use some work. However, it's always
 *         going to be heuristic) 
 *
 *   Arguments:
 *   
 *    fp           = File pointer
 *    nTitleLines = Number of title lines
 *    nColTitleLines = Number of column title lines
 *    nCol        = Number of columns -> basically equal to the 
 *                  number of variabless
 *    nDataRows   = Number of rows of data in the file
 *
 */

static void get_sizes(FILE *fp, int &nTitleLines, int &nColTitleLines, 
	              int &nCol, int &nDataRows, int **ColIsFloat_ptr)
{
  int nScanLinesMAX = 100;
  int nScanLines = nScanLinesMAX;
  int retn, i, j;
  int maxCommas = 0;
  TOKEN fieldToken;
  char *scanLine = mdp_alloc_char_1(MAX_INPUT_STR_LN+1, '\0');
  int *numCommas = mdp_alloc_int_1(nScanLinesMAX, -1);
  int *ColIsFloat = *ColIsFloat_ptr;

  /*
   * Rewind the file
   */
  rewind(fp);
  /*
   * Read the scan lines
   */
  for (i = 0; i < nScanLinesMAX; i++) {
    retn = read_line(fp, scanLine, 0);
    if (retn == -1) {
      nScanLines = i;
      break;
    }
    /*
     * Strip a trailing comma from the scanline -
     *  -> These are not significant
     */
    int ccount = static_cast<int>(strlen(scanLine));
    if (ccount > 0) {
      if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
    }
    /*
     * Count the number of commas in the line
     */
    char *cptr = scanLine;
    char *cetn = NULL;
    numCommas[i] = 0;
    do {
      cetn = strchr(cptr, (int) ',');
      if (cetn) {
	numCommas[i]++;
	cptr = cetn + 1;
      }
    } while (cetn);
    if (i > 1) {
      if ( maxCommas < numCommas[i]) {
	maxCommas = numCommas[i];
      }
    }
  }
  /*
   * set a preliminary value of nCol
   */
  nCol = maxCommas + 1;
  if (nScanLines == 0) {
    nCol = 0;
  }
  char **strlets = (char **) mdp_alloc_ptr_1(maxCommas+1);

  /*
   * Figure out if each column is a text or float
   */
  rewind(fp);
  for (i = 0; i < nScanLines; i++) {
    retn = read_line(fp, scanLine, 0);
    int ncolsFound = breakStrCommas(scanLine, strlets, nCol);
    if (ncolsFound == (maxCommas + 1)) {
      for (j = 0; j < ncolsFound; j++) {
	char *fieldStr = strlets[j];
	fillTokStruct(&fieldToken, fieldStr);
	if (fieldToken.ntokes != 1) {
	  break;
	}
	bool rerr = false;
	(void) tok_to_double(&fieldToken, DBL_MAX,
			     -DBL_MAX, 0.0, &rerr);
	if (!rerr) {
	  ColIsFloat[j] = true;
	}
      }

    }
  }



  int doingLineType = LT_TITLELINE;
  rewind(fp);
  for (i = 0; i < nScanLines; i++) {
    retn = read_line(fp, scanLine, 0);
    /*
     * Strip a trailing comma from the scanline -
     *  -> These are not significant
     */
    int ccount = static_cast<int>(strlen(scanLine));
    if (ccount > 0) {
      if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
    }
    int ncolsFound = breakStrCommas(scanLine, strlets, nCol);

    if (doingLineType == LT_TITLELINE) {
      if (numCommas[i] == maxCommas) {
	doingLineType = LT_COLTITLE;
	nTitleLines = i;
      }
    }

    if (doingLineType == LT_COLTITLE) {
      bool goodDataLine = true;
      bool rerr = false;
      for (j = 0; j < ncolsFound; j++) {
	char *fieldStr = strlets[j];
	fillTokStruct(&fieldToken, fieldStr);
	if (fieldToken.ntokes != 1) {
	  goodDataLine = false;
	  break;
	}
	if ((ColIsFloat[j]) == 1) {
	  (void) tok_to_double(&fieldToken, DBL_MAX,
			       -DBL_MAX, 0.0, &rerr);
	  if (rerr) {
	    goodDataLine = false;
	    break;
	  }
	}
      }
      if (goodDataLine) {
	doingLineType = LT_DATALINE;
      }
      nColTitleLines = i - nTitleLines;
    }
    if (doingLineType == LT_DATALINE) break;
  }


  /*
   * Count the total number of lines in the file
   */
  if (doingLineType == LT_DATALINE) {
    for (i = nColTitleLines + nTitleLines; ; i++) {
      retn = read_line(fp, scanLine, 0);
      if (retn == -1) {
	nDataRows = i - nColTitleLines - nTitleLines + 1;
	break;
      }
      /*
       * Strip a trailing comma from the scanline -
       *  -> These are not significant
       */
      int ccount = static_cast<int>(strlen(scanLine));
      if (ccount > 0) {
	if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
      }
      int ncolsFound = breakStrCommas(scanLine, strlets, nCol);   
      bool goodDataLine = true;
      bool rerr = false;
      for (j = 0; j < ncolsFound; j++) {
	char *fieldStr = strlets[j];
	fillTokStruct(&fieldToken, fieldStr);
	if (fieldToken.ntokes != 1) {
	  goodDataLine = false;
	  break;
	}
	if (ColIsFloat[j] == 1) {
	  (void) tok_to_double(&fieldToken, DBL_MAX,
			       -DBL_MAX, 0.0, &rerr);
	  if (rerr) {
	    goodDataLine = false;
	    break;
	  }
	}
      }
      if (! goodDataLine) {
	doingLineType = LT_NULLLINE;
	nDataRows = i - nColTitleLines - nTitleLines + 1;
	break;
      }
    }
  }
  mdp_safe_free((void **) &strlets);
  mdp_safe_free((void **) &scanLine);
  mdp_safe_free((void **) &numCommas);
  return;
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static void
read_title(FILE *fp, char ***title, int nTitleLines)
{
  int retn;
  *title = (char **) mdp_alloc_ptr_1(nTitleLines);
  char *scanLine = mdp_alloc_char_1(Max_Input_Str_Ln + 1, '\0');
   for (int i = 0; i < nTitleLines ; i++) {
      retn = read_line(fp, scanLine, 0);
      if (retn >= 0) {
	/*
	 * Strip a trailing comma from the scanline -
	 *  -> These are not significant
	 */
	int ccount = static_cast<int>(strlen(scanLine));
	if (ccount > 0) {
	  if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
	}
	(*title)[i] = mdp_copy_string(scanLine);
      }
   }
   mdp_safe_free((void **) &scanLine);
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static void
read_colTitle(FILE *fp,  char ****ColMLNames_ptr, int  nColTitleLines, int nCol)
{
  int retn, j;
  *ColMLNames_ptr = (char ***) mdp_alloc_ptr_1(nCol);
  char ***ColMLNames = *ColMLNames_ptr;
  char *scanLine = mdp_alloc_char_1(Max_Input_Str_Ln + 1, '\0');
  char **strlets = (char **) mdp_alloc_ptr_1(nCol+1);
  if (nColTitleLines > 0) {
    for (int i = 0; i < nColTitleLines ; i++) {
      retn = read_line(fp, scanLine, 0);
      if (retn >= 0) {
	/*
	 * Strip a trailing comma from the scanline -
	 *  -> These are not significant
	 */
	int ccount = static_cast<int>(strlen(scanLine));
	if (ccount > 0) {
	  if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
	}
	int ncolsFound = breakStrCommas(scanLine, strlets, nCol);
	ColMLNames[i] = mdp_alloc_VecFixedStrings(nCol, MAX_TOKEN_STR_LN+1);
	for (j = 0; j < ncolsFound; j++) {
	  strip(strlets[j]);
	  strcpy(ColMLNames[i][j], strlets[j]);
	}
      }
    }
  } else {
    ColMLNames[0] = mdp_alloc_VecFixedStrings(nCol, MAX_TOKEN_STR_LN+1);
    for (j = 0; j < nCol; j++) {
      char cbuff[256];
      sprintf(cbuff, "Col_%d", j+1);
      strcpy(ColMLNames[0][j], cbuff);
    }
  }
  mdp_safe_free((void **) &scanLine);
  mdp_safe_free((void **) &strlets);
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static double get_atol(const double *values, const int nvals,
		       const double atol)
{
 int i;
 double sum = 0.0, retn;
 if (nvals <= 0) return gatol;
 for (i = 0; i < nvals; i++) {
   retn = values[i];
   sum += retn * retn;
 }
 sum /= nvals;
 retn = sqrt(sum);
 return ((retn + 1.0) * atol);
}

/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static void
read_values(FILE *fp, double **NVValues, char ***NSValues, int nCol, int nDataRows,
	    int *ColIsFloat)
{
  char **strlets = (char **) mdp_alloc_ptr_1(nCol+1);
  char *scanLine = mdp_alloc_char_1(Max_Input_Str_Ln + 1, '\0');
  TOKEN fieldToken;
  double value;
  int retn, j;
  for (int i = 0; i < nDataRows; i++) {
    retn = read_line(fp, scanLine, 0);
    if (retn == -1) {
      break;
    }
    /*
     * Strip a trailing comma from the scanline -
     *  -> These are not significant
     */
    int ccount = static_cast<int>(strlen(scanLine));
    if (ccount > 0) {
      if (scanLine[ccount-1] == ',') scanLine[ccount-1] = '\0';
    }
    int ncolsFound = breakStrCommas(scanLine, strlets, nCol);   
    bool goodDataLine = true;
    bool rerr = false;
    for (j = 0; j < ncolsFound; j++) {
      char *fieldStr = strlets[j];
      NSValues[j][i] = mdp_copy_string(strlets[j]);
      fillTokStruct(&fieldToken, fieldStr);
      if (fieldToken.ntokes != 1) {
	goodDataLine = false;
	break;
      }
      if (ColIsFloat[j]) {
	value = tok_to_double(&fieldToken, DBL_MAX,
			      -DBL_MAX, 0.0, &rerr);
	if (rerr) {
	  goodDataLine = false;
	  break;
	}
	NVValues[j][i] = value;
      }
    }
    if (! goodDataLine) {
      break;
    }
  }
  mdp_safe_free((void **) &strlets);
  mdp_safe_free((void **) &scanLine);
}
/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

static void print_usage() {
  printf("\t\n");
  printf("  csvdiff [-h] [-a atol] [-r rtol] File1.csv File2.csv\n");
  printf("\t\n");
  printf("\tCompares the variable values in two Excel formatted "
         "comma separated files.\n");
  printf("\tThe comparison is done using a weighted norm basis.\n");
  printf("\t\n");
  printf("\tThe two files should be basically equal. However, File1.csv is\n");
  printf("\ttaken as the reference file that has precedence, when there is\n");
  printf("\tsomething to be decided upon.\n");
  printf("\t\n");
  printf("\t Arguments:\n");
  printf("\t  -h      = Usage info\n");
  printf("\t  -a atol = Set absolute tolerance parameter - default = 1.0E-9\n");
  printf("\t  -r rtol = Set relative tolerance parameter - default = 1.0E-3\n");
  printf("\t\n");
  printf("\t Shell Return Values:\n");
  printf("\t   0 = Comparison was successful\n");
  printf("\t   1 = One or more nodal values failed the comparison\n");
  printf("\t   2 = One or more header values failed the comparison\n");
  printf("\t   3 = Apples to oranges, the files can not even be compared against\n");
  printf("\t       one another.\n");
  printf("\t\n");
}
/*****************************************************************************/
/*****************************************************************************/
/*****************************************************************************/

int main(int argc, char *argv[])
  
  /*
   *                        main driver for csvdiff.
   */
{
  int opt_let;
  char  *fileName1=NULL, *fileName2=NULL; /* Names of the csv files */
  FILE  *fp1=NULL, *fp2=NULL;
  int    nTitleLines1 = 0, nTitleLines2 = 0;
  int    nColTitleLines1 = 0, nColTitleLines2 = 0;
  int    nCol1 = 0,  nCol2 = 0, nColMAX = 0, nColcomparisons = 0;
  int    nDataRows1 = 0, nDataRows2 = 0;
  char  **title1 = 0,  **title2 = 0;
  int    **compColList = NULL;
  char   ***ColMLNames1 = NULL, ***ColMLNames2 = NULL;
  char   **ColNames1 = NULL, **ColNames2 = NULL;
  double **NVValues1 = NULL, **NVValues2 = NULL;
  char   ***NSValues1 = NULL, ***NSValues2 = NULL; 
  int    *ColIsFloat1 = NULL, *ColIsFloat2 = NULL;
  double *curVarValues1 = NULL, *curVarValues2 = NULL;
  char ** curStringValues1 = NULL, **curStringValues2 = NULL;
  int    i, j, ndiff, jmax, i1, i2, k;
  bool found;
  double max_diff, rel_diff;
  int    testPassed = RT_PASSED;
  double atol_j, atol_arg = 0.0, rtol_arg = 0.0;
  
  /********************** BEGIN EXECUTION ************************************/
  int id = 0;
  int id2 = 0;
  char *ggg = 0;
  char *rrr = 0;
  /*
  * Interpret command line arguments
  */
    /* Loop over each command line option */
  while((opt_let = getopt(argc, argv, "ha:r:")) != EOF) {

    /* case over the option letter */
    switch(opt_let) {
      
    case 'h':
      /* Usage info was requested */
      print_usage();
      exit(0);

   case 'a':
     /* atol parameter */
 
     ggg = optarg;
     //printf("a = %s\n", ggg);
     id = sscanf(ggg,"%lg", &atol_arg);
     if (id != 1) {
       printf(" atol param bad: %s\n", ggg);
       exit(-1);
     }
     gatol = atol_arg;
     break;

   case 'r':
     /* rtol parameter */
 
     rrr = optarg;
     //printf("r = %s\n", ggg);
     id2 = sscanf(rrr,"%lg", &rtol_arg);
     if (id2 != 1) {
       printf(" rtol param bad: %s\n", rrr);
       exit(-1);
     }
     grtol = rtol_arg;
     break;
    
      
    default:
      /* Default case. Error on unknown argument. */
      printf("default called opt_let = %c\n", opt_let);
      fprintf(stderr, "ERROR in command line usuage:\n");
      print_usage();
      return 0; 
    } /* End "switch(opt_let)" */

  } /* End "while((opt_let=getopt(argc, argv, "i")) != EOF)" */

  if (optind !=  argc-2) {
    print_usage();
    exit(-1);
  } else {
    fileName1 = argv[argc-2];
    fileName2 = argv[argc-1];
  }

  /*
   *      Print Out Header
   */
  printf("\n");
  printf("----------------------------------------------------------\n");
  printf("csvdiff: CSVFile comparison utility program\n");
  printf("         Version $Revision$\n");
  printf("         Harry K. Moffat Div. 9114 Sandia National Labs\n");
  printf("         \n");
  printf("         First  CSV File = %s\n", fileName1);
  printf("         Second CSV file = %s\n", fileName2); 
  printf("\n");
  printf("         Absolute tol = %g\n", gatol);
  printf("         Relative tol = %g\n", grtol);
  printf("----------------------------------------------------------\n");
  printf("\n");  
  
  /*
  *  Open up the two ascii Files #1 and #2
  */
  if (!(fp1 = fopen(fileName1, "r"))) {
    fprintf(stderr,"Error opening up file1, %s\n", fileName1);
    exit(-1);
  }
  if (!(fp2 = fopen(fileName2, "r"))) {
    fprintf(stderr, "Error opening up file2, %s\n", fileName2);
    exit(-1);
  }

  ColIsFloat1 = mdp_alloc_int_1(200, 0);
  ColIsFloat2 = mdp_alloc_int_1(200, 0);
  /*
   *   Obtain the size of the problem information: Compare between files. 
   */

  get_sizes(fp1, nTitleLines1, nColTitleLines1, nCol1, nDataRows1, &ColIsFloat1);
  if (nCol1 == 0) {
    printf("Number of columns in file %s is zero\n", fileName1);
    testPassed = RT_FAILED_OTHER;
    exit(RT_FAILED_OTHER);
  }
  if (nDataRows1 == 0) {
    printf("Number of data rows in file %s is zero\n", fileName1);
    testPassed = RT_FAILED_OTHER;
    exit(RT_FAILED_OTHER);
  }

  get_sizes(fp2, nTitleLines2, nColTitleLines2, nCol2, nDataRows2, &ColIsFloat2);
  if (nCol2 == 0) {
    printf("Number of columns in file %s is zero\n", fileName2);
    testPassed = RT_FAILED_OTHER;
    exit(RT_FAILED_OTHER);
  }
  if (nDataRows2 == 0) {
    printf("Number of data rows in file %s is zero\n", fileName2);
    testPassed = RT_FAILED_OTHER;
    exit(RT_FAILED_OTHER);
  }

  if (nTitleLines1 != nTitleLines2) {
    printf("Number of Title Lines differ:, %d %d\n",nTitleLines1, nTitleLines2);
    testPassed = RT_FAILED_OTHER;
  } else if (Debug_Flag) {
    printf("Number of Title Lines in each file = %d\n", nTitleLines1);
  }
  if (nColTitleLines1 != nColTitleLines2) {
    printf("Number of Column title lines differ:, %d %d\n", nColTitleLines1,
	   nColTitleLines2);
    testPassed = RT_FAILED_OTHER;
  } else if (Debug_Flag) {
    printf("Number of column title lines in each file = %d\n", nColTitleLines1);
  }

  /*
   * Right now, if the number of data rows differ, we will punt.
   * Maybe later we can do something more significant
   */
  int nDataRowsMIN = MIN(nDataRows1, nDataRows2);
  int nDataRowsMAX = MAX(nDataRows1, nDataRows2);
  if (nDataRows1 != nDataRows2) {
    printf("Number of Data rows in file1, %d, is different than file2, %d\n",
	   nDataRows1, nDataRows2);
  } else {
    printf("Number of Data rows in both files = %d\n", nDataRowsMIN); 
  }

  rewind(fp1);
  rewind(fp2);
  read_title(fp1, &title1, nTitleLines1);
  read_title(fp2, &title2, nTitleLines2);

  if (nTitleLines1 > 0 || nTitleLines2 > 0) {
    int n = MIN(nTitleLines1, nTitleLines2);
    for (i = 0; i < n; i++) {
      if (strcmp(title1[i], title2[i]) != 0) {
	printf("Title Line %d differ:\n\t\"%s\"\n\t\"%s\"\n", i, title1[i], title2[i]);
        testPassed = RT_FAILED_HDR;
      } else if (Debug_Flag) {
	printf("Title Line %d for each file: \"%s\"\n", i, title1[i]);
      }
    }
    if (nTitleLines1 != nTitleLines2) {
      printf("Number of Title Lines differ: %d %d\n", nTitleLines1, nTitleLines2);
      testPassed = RT_FAILED_HDR;;
    }
  } else {
    if (nTitleLines1 != nTitleLines2) {
      if (nTitleLines1) {
	printf("Titles differ: title for first file: \"%s\"\n", 
	       title1[0]);
        testPassed = RT_FAILED_HDR;
      }
      if (nTitleLines2) {
	printf("Titles differ: title for second file: \"%s\"\n", 
	       title2[0]);
      }
      testPassed = RT_FAILED_HDR;
    }
  }

  /*
   *  Get the number of column variables in each file
   */

  if (nCol1 != nCol2) {
    printf("Number of column variables differ:, %d %d\n", 
                 nCol1, nCol2);
    testPassed = RT_FAILED_OTHER;
  } else if (Debug_Flag) {
    printf("Number of column variables in both files = %d\n",
	   nCol1);
  }

  /*
   * Read the names of the column variables
   */
  read_colTitle(fp1, &ColMLNames1,  nColTitleLines1, nCol1);
  read_colTitle(fp2, &ColMLNames2,  nColTitleLines2, nCol2);
  ColNames1 = ColMLNames1[0];
  ColNames2 = ColMLNames2[0];

  /*
   * Do a Comparison of the names to find the maximum number
   * of matches.
   */
  nColMAX = MAX(nCol1, nCol2);

  compColList = mdp_alloc_int_2(nColMAX, 2, -1);
  nColcomparisons = 0;
  for (i = 0; i < nCol1; i++) {
    found = false;
    for (j = 0; j < nCol2; j++) {
      if (!strcmp(ColNames1[i], ColNames2[j])) {
	compColList[nColcomparisons][0] = i;
	compColList[nColcomparisons][1] = j;
	nColcomparisons++;
	found = true;
	break;
      }
    }
    if (!found) {
      printf("csvdiff WARNING Variable %s (%d) in first file not found"
	     " in second file\n", ColNames1[i], i);
      testPassed = RT_FAILED_OTHER;
    }
  }
  for (j = 0; j < nCol2; j++) {
    found = false;
    for (i = 0; i <  nColcomparisons; i++) {
       if (compColList[i][1] == j) found = true;
    }
    if (! found) {
      printf("csvdiff WARNING Variable %s (%d) in second file "
	     "not found in first file\n",
	     ColNames2[j], j);
      testPassed = RT_FAILED_OTHER;
    }
  }

  /*
   *  Allocate storage for the column variables
   */
  NVValues1 = mdp_alloc_dbl_2(nCol1, nDataRowsMAX, 0.0);
  NVValues2 = mdp_alloc_dbl_2(nCol2, nDataRowsMAX, 0.0);

  /*
   *  Allocate storage for the column variables
   */
  NSValues1 = (char ***) mdp_alloc_ptr_2(nCol1, nDataRowsMAX);
  NSValues2 = (char ***) mdp_alloc_ptr_2(nCol2, nDataRowsMAX);

  /*
   *  Read in the values to the arrays
   */
  read_values(fp1, NVValues1, NSValues1, nCol1, nDataRows1, ColIsFloat1);
  read_values(fp2, NVValues2, NSValues2, nCol2, nDataRows2, ColIsFloat2);

  /*
   * Compare the solutions in each file
   */
  int method = 0;
#define DGG_MODS
#ifdef DGG_MODS
  method = 1;
#endif
  double slope1, slope2, xatol;
  int notOK;
  for (k = 0; k < nColcomparisons; k++) {

    i1 =  compColList[k][0];
    i2 =  compColList[k][1];
    bool doFltComparison = true;
    if (!ColIsFloat1[i1]) {
      doFltComparison = false;
      jmax = -1;
    }
    if (!ColIsFloat2[i2]) {
      doFltComparison = false;
      jmax = -1;
    }
    curStringValues1 = NSValues1[i1];
    curStringValues2 = NSValues2[i2]; 
    max_diff = 0.0;
    ndiff = 0;
    if (doFltComparison) {
      curVarValues1 = NVValues1[i1];
      curVarValues2 = NVValues2[i2];
      atol_j =             get_atol(curVarValues1, nDataRows1, gatol);
      atol_j = MIN(atol_j, get_atol(curVarValues2, nDataRows2, gatol));
      for (j = 0; j < nDataRowsMIN; j++) {

	slope1 = 0.0;
	slope2 = 0.0;
	xatol = fabs(grtol * (NVValues1[0][j] - NVValues1[0][j-1]));
	if (j > 0 && k > 0) {
	  slope1 = (curVarValues1[j] - curVarValues1[j-1])/
	    (NVValues1[0][j] - NVValues1[0][j-1]);
	  slope2 = (curVarValues2[j] - curVarValues2[j-1])/
	    (NVValues2[0][j] - NVValues2[0][j-1]);
	}
	if (method) {
	  notOK = diff_double_slope(curVarValues1[j], curVarValues2[j], 
				    grtol, atol_j, xatol, slope1, slope2);
	} else {
	  notOK = diff_double(curVarValues1[j], curVarValues2[j], 
			      grtol, atol_j);
	}
	if (notOK) {
	  ndiff++;
	  rel_diff = calc_rdiff((double) curVarValues1[j], 
				(double) curVarValues2[j], grtol, atol_j);
	  if (rel_diff > max_diff) {
	    jmax = j;
	    max_diff = rel_diff;
	  }
	  if (ndiff < 10) {
	    printf("\tColumn variable %s at data row %d ", ColNames1[i1], j + 1);
	    printf(" differ: %g %g\n", curVarValues1[j], 
		   curVarValues2[j]);
	  }
	}
      }
    } else {
      for (j = 0; j < nDataRowsMIN; j++) {
	strip(curStringValues1[j]);
	strip(curStringValues2[j]);
	notOK = false;
	if (strcmp(curStringValues1[j], curStringValues2[j])) {
	  notOK = true;
	  ndiff++;
	  if (ndiff < 10) {
	    printf("\tColumn String variable %s at data row %d ", ColNames1[i1], j + 1);
	    printf(" differ: %s %s\n", curStringValues1[j], 
		   curStringValues2[j]);
	  }
	}
      }
    }
    
    if (nDataRowsMIN != nDataRowsMAX) {
      ndiff +=  nDataRowsMAX - nDataRowsMIN;
      if (ndiff < 10) {
	if (nDataRows1 > nDataRows2) {
	  for (j = nDataRowsMIN; j < nDataRowsMAX; j++) {
	    printf("\tColumn variable %s at data row %d ", ColNames1[i1], j + 1);
	    printf(" differ: %g      NA\n", curVarValues1[j]);
	  }
	} else {
	  for (j = nDataRowsMIN; j < nDataRowsMAX; j++) {
	    printf("\tColumn variable %s at data row %d ", ColNames1[i1], j + 1);
	    printf(" differ: NA     %g \n", curVarValues2[j]);
	  }
	}
      }
    }

    /*
     *  Print out final results of nodal variable test
     */

    if (ndiff > 0) {
      printf(
	     "Column variable %s failed comparison test for %d occurances\n",
	     ColNames1[i1], ndiff);
      if (jmax >= 0) {
	printf("  Largest difference was at data row %d ", jmax + 1);
	printf(": %g %g\n", curVarValues1[jmax],  curVarValues2[jmax]);
      }
      testPassed = RT_FAILED_COL;
    } else if (Debug_Flag) {
      printf("Column variable %s passed\n",  ColNames1[i1]);
    }
 
  }
  
  return(testPassed);

} /************END of main() *************************************************/
/*****************************************************************************/