6 files changed, 293 insertions, 0 deletions
diff --git a/src/s3select/example/CMakeLists.txt b/src/s3select/example/CMakeLists.txt
new file mode 100644
index 000000000..37f99f0b3
--- /dev/null
+++ b/src/s3select/example/CMakeLists.txt
@@ -0,0 +1,12 @@
+add_executable(s3select_example s3select_example.cpp)
+target_include_directories(s3select_example PUBLIC ../include)
+target_link_libraries(s3select_example boost_date_time)
+
+add_executable(generate_rand_csv generate_rand_csv.c)
+
+add_custom_command(OUTPUT expr_genrator.py COMMAND cp ${CMAKE_CURRENT_SOURCE_DIR}/expr_genrator.py expr_genrator.py
+    COMMENT "Copy expr_genrator.py"
+    VERBATIM)
+
+add_custom_target(expr_generator ALL DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/expr_genrator.py)
+
diff --git a/src/s3select/example/expr_genrator.py b/src/s3select/example/expr_genrator.py
new file mode 100755
index 000000000..0d21fcee6
--- /dev/null
+++ b/src/s3select/example/expr_genrator.py
@@ -0,0 +1,9 @@
+import random
+import sys
+
+def expr(depth):
+    if depth==1 or random.random()<1.0/(2**depth-1): 
+        return str(int(random.random() * 100) + 1)+".0"
+    return '(' + expr(depth-1) + random.choice(['+','-','*','/']) + expr(depth-1) + ')'
+
+print expr( int(sys.argv[1]) )
diff --git a/src/s3select/example/generate_rand_csv.c b/src/s3select/example/generate_rand_csv.c
new file mode 100644
index 000000000..67d52adaa
--- /dev/null
+++ b/src/s3select/example/generate_rand_csv.c
@@ -0,0 +1,28 @@
+#include <stdio.h>
+#include <stdlib.h>
+
+
+int main(int argc, char** argv)
+{
+  if (argc<3)
+  {
+    printf("%s <num-of-rows> <num-of-columns> \n", argv[0]);
+    return -1;
+  }
+
+  srand(1234);
+  int line_no=0;
+  for(int i=0; i<atoi(argv[1]); i++)
+  {
+    printf("%d,", i);
+    for(int y=0; y<atoi(argv[2]); y++)
+    {
+      printf("%d,", rand()%1000);
+    }
+    printf("\n");
+  }
+
+
+
+
+}
diff --git a/src/s3select/example/parse_csv.py b/src/s3select/example/parse_csv.py
new file mode 100755
index 000000000..23fe14add
--- /dev/null
+++ b/src/s3select/example/parse_csv.py
@@ -0,0 +1,12 @@
+#!/usr/bin/python
+
+import csv
+
+with open('stam.txt') as csv_file:
+    csv_reader = csv.reader(csv_file, delimiter=',')
+    line_count = 0
+    for row in csv_reader:
+        #if (int(row[0])==465 and int(row[5])==268): # casting is slower
+        if (row[0]=="465" and row[5]=="268"):
+            print row
+
diff --git a/src/s3select/example/run_test.bash b/src/s3select/example/run_test.bash
new file mode 100755
index 000000000..56654d7ab
--- /dev/null
+++ b/src/s3select/example/run_test.bash
@@ -0,0 +1,96 @@
+#!/bin/bash
+
+set -e
+
+PREFIX=${1:-"./example"} 
+
+## purpose : sanity tests
+
+s3select_calc() 
+{
+l="$*"  
+res=$( echo 1 | "$PREFIX"/s3select_example -q  "select ${l} from stdin;" ) 
+echo "$res" | sed 's/.$//'
+}
+
+# create c file with expression , compile it and run it.
+c_calc()
+{
+cat << @@ > "$PREFIX"/tmp.c
+
+#include <stdio.h>
+int main()
+{
+printf("%f\n",$*);
+}
+@@
+gcc -o "$PREFIX"/a.out "$PREFIX"/tmp.c
+"$PREFIX"/a.out
+}
+
+expr_test()
+{
+## test the arithmetic evaluation of s3select against C program 
+for i in {1..100}
+do
+	e=$(python2 "$PREFIX"/expr_genrator.py 5)
+	echo expression["$i"]="$e"
+	r1=$(s3select_calc "$e")
+	r2=$(c_calc "$e")
+    echo "$r1" "$r2"
+
+	## should be zero or very close to zero; ( s3select is C compile program )
+    res=$(echo "" | awk -v e="$e" -v r1="$r1" -v r2="$r2" 'function abs(n){if (n<0) return -n; else return n;}{if (abs(r1-r2) > 0.00001) {print "MISSMATCH result for expression",e;}}')
+    if test "$res" != ""; then
+        echo "$res"
+        exit 1
+    fi
+done
+}
+
+aggregate_test()
+{
+## generate_rand_csv is generating with the same seed 
+echo check sum 
+s3select_val=$("$PREFIX"/generate_rand_csv 10 10 | "$PREFIX"/s3select_example -q 'select sum(int(_1)) from stdin;') 
+awk_val=$("$PREFIX"/generate_rand_csv 10 10 | awk 'BEGIN{FS=",";} {s+=$1;} END{print s;}')
+s3select_val=${s3select_val::-1}
+echo "$s3select_val" "$awk_val"
+if test "$s3select_val" -ne "$awk_val"; then
+    exit 1
+fi
+echo check min 
+s3select_val=$("$PREFIX"/generate_rand_csv 10 10 | "$PREFIX"/s3select_example -q 'select min(int(_1)) from stdin;') 
+awk_val=$("$PREFIX"/generate_rand_csv 10 10 | awk 'BEGIN{FS=",";min=100000;} {if(min>$1) min=$1;} END{print min;}')
+s3select_val=${s3select_val::-1}
+echo "$s3select_val" "$awk_val"
+if test "$s3select_val" -ne "$awk_val"; then
+    exit 1
+fi
+echo check max 
+s3select_val=$("$PREFIX"/generate_rand_csv 10 10 | "$PREFIX"/s3select_example -q 'select max(int(_1)) from stdin;') 
+awk_val=$("$PREFIX"/generate_rand_csv 10 10 | awk 'BEGIN{FS=",";max=0;} {if(max<$1) max=$1;} END{print max;}' )
+s3select_val=${s3select_val::-1}
+echo "$s3select_val" "$awk_val"
+if test "$s3select_val" -ne "$awk_val"; then
+    exit 1
+fi
+echo check substr and count 
+s3select_val=$("$PREFIX"/generate_rand_csv 10000 10 | "$PREFIX"/s3select_example -q 'select count(int(_1)) from stdin where int(_1)>200 and int(_1)<250;')
+awk_val=$("$PREFIX"/generate_rand_csv 10000 10 | "$PREFIX"/s3select_example -q 'select substr(_1,1,1) from stdin where int(_1)>200 and int(_1)<250;' | uniq -c | awk '{print $1;}')
+s3select_val=${s3select_val::-1}
+echo "$s3select_val" "$awk_val"
+if test "$s3select_val" -ne "$awk_val"; then
+    exit 1
+fi
+}
+
+###############################################################
+
+expr_test
+aggregate_test
+
+rm "$PREFIX"/tmp.c "$PREFIX"/a.out
+
+exit 0
+
diff --git a/src/s3select/example/s3select_example.cpp b/src/s3select/example/s3select_example.cpp
new file mode 100644
index 000000000..840b62c6a
--- /dev/null
+++ b/src/s3select/example/s3select_example.cpp
@@ -0,0 +1,136 @@
+#include "s3select.h"
+#include <fstream>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+using namespace s3selectEngine;
+using namespace BOOST_SPIRIT_CLASSIC_NS;
+
+int cli_get_schema(const char* input_schema, actionQ& x)
+{
+  g_push_column.set_action_q(&x);
+
+  rule<> column_name_rule = lexeme_d[(+alpha_p >> *digit_p)];
+
+  //TODO an issue to resolve with trailing space
+  parse_info<> info = parse(input_schema, ((column_name_rule)[BOOST_BIND_ACTION(push_column)] >> *(',' >> (column_name_rule)[BOOST_BIND_ACTION(push_column)])), space_p);
+
+  if (!info.full)
+  {
+    std::cout << "failure in schema description " << input_schema << std::endl;
+    return -1;
+  }
+
+  return 0;
+}
+
+int main(int argc, char** argv)
+{
+
+  //purpose: demostrate the s3select functionalities
+  s3select s3select_syntax;
+
+  char*  input_query = 0;
+
+  for (int i = 0; i < argc; i++)
+  {
+
+    if (!strcmp(argv[i], "-q"))
+    {
+      input_query = argv[i + 1];
+    }
+  }
+
+
+  if (!input_query)
+  {
+    std::cout << "type -q 'select ... from ...  '" << std::endl;
+    return -1;
+  }
+
+
+  bool to_aggregate = false;
+
+  int status = s3select_syntax.parse_query(input_query);
+  if (status != 0)
+  {
+    std::cout << "failed to parse query " << s3select_syntax.get_error_description() << std::endl;
+    return -1;
+  }
+
+  std::string object_name = s3select_syntax.get_from_clause(); //TODO stdin
+
+  FILE* fp;
+
+  if (object_name.compare("stdin")==0)
+  {
+    fp = stdin;
+  }
+  else
+  {
+    fp  = fopen(object_name.c_str(), "r");
+  }
+
+
+  if(!fp)
+  {
+    std::cout << " input stream is not valid, abort;" << std::endl;
+    return -1;
+  }
+
+  struct stat statbuf;
+
+  lstat(object_name.c_str(), &statbuf);
+
+  std::string s3select_result;
+  s3selectEngine::csv_object::csv_defintions csv;
+  csv.use_header_info = false;
+  //csv.column_delimiter='|';
+  //csv.row_delimiter='\t';
+
+
+  s3selectEngine::csv_object  s3_csv_object(&s3select_syntax, csv);
+  //s3selectEngine::csv_object  s3_csv_object(&s3select_syntax);
+
+#define BUFF_SIZE 1024*1024*4
+  char* buff = (char*)malloc( BUFF_SIZE );
+  while(1)
+  {
+    //char buff[4096];
+
+    //char * in = fgets(buff,sizeof(buff),fp);
+    size_t input_sz = fread(buff, 1, BUFF_SIZE, fp);
+    char* in=buff;
+    //input_sz = strlen(buff);
+    //size_t input_sz = in == 0 ? 0 : strlen(in);
+
+    //if (!input_sz) to_aggregate = true;
+
+
+    //int status = s3_csv_object.run_s3select_on_object(s3select_result,in,input_sz,false,false,to_aggregate);
+    int status = s3_csv_object.run_s3select_on_stream(s3select_result, in, input_sz, statbuf.st_size);
+    if(status<0)
+    {
+      std::cout << "failure on execution " << std::endl;
+      break;
+    }
+
+    if(s3select_result.size()>1)
+    {
+      std::cout << s3select_result;
+    }
+
+    s3select_result = "";
+    if(!input_sz || feof(fp))
+    {
+      break;
+    }
+
+  }
+
+  free(buff);
+  fclose(fp);
+
+
+}