SQLite

  /* There is a quirk here. The users INSERT statement may have specified
  ** a value for the "rowid" field, for the "docid" field, or for both.
  ** Which is a problem, since "rowid" and "docid" are aliases for the
  ** same value. For example:
  **
  **   INSERT INTO fts3tbl(rowid, docid) VALUES(1, 2);
  **
  ** In FTS3, this is an error. It is an error to specify non-NULL values
  ** for both docid and some other rowid alias.
  */
  if( SQLITE_NULL!=sqlite3_value_type(apVal[3+p->nColumn]) ){
    if( SQLITE_NULL==sqlite3_value_type(apVal[0])
     && SQLITE_NULL!=sqlite3_value_type(apVal[1])
    ){
      /* A rowid/docid conflict. */
      return SQLITE_ERROR;
    }
    rc = sqlite3_bind_value(pContentInsert, 1, apVal[3+p->nColumn]);
    if( rc!=SQLITE_OK ) return rc;
  }

  /* Execute the statement to insert the record. Set *piDocid to the 
  ** new docid value. 
  */

  int nAlloc = 0;                 /* Allocated size of aBuffer buffer */
  int rc = SQLITE_OK;             /* Return code */

  int isIgnoreEmpty =  (pFilter->flags & FTS3_SEGMENT_IGNORE_EMPTY);
  int isRequirePos =   (pFilter->flags & FTS3_SEGMENT_REQUIRE_POS);
  int isColFilter =    (pFilter->flags & FTS3_SEGMENT_COLUMN_FILTER);
  int isPrefix =       (pFilter->flags & FTS3_SEGMENT_PREFIX);

  /* If there are zero segments, this function is a no-op. This scenario
  ** comes about only when reading from an empty database.
  */
  if( nSegment==0 ) goto finished;

  /* If the Fts3SegFilter defines a specific term (or term prefix) to search 
  ** for, then advance each segment iterator until it points to a term of
  ** equal or greater value than the specified term. This prevents many
  ** unnecessary merge/sort operations for the case where single segment
  ** b-tree leaf nodes contain more than one term.
  */

# This file implements tests to verify the "testable statements" in the
# fts3.in document.
#

set testdir [file dirname $argv0]
source $testdir/tester.tcl

# If this build does not include FTS3, skip the tests in this file.
#
ifcapable !fts3 { finish_test ; return }



source $testdir/fts3_common.tcl

# Procs used to make the tests in this file easier to read.
#
proc ddl_test {tn ddl} {
  uplevel [list do_write_test e_fts3-$tn sqlite_master $ddl]
}
proc write_test {tn tbl sql} {
  uplevel [list do_write_test e_fts3-$tn $tbl $sql]
}
proc read_test {tn sql result} {
  uplevel [list do_select_test e_fts3-$tn $sql $result]
}


#-------------------------------------------------------------------------
# The body of the following [foreach] block contains test cases to verify
# that the example code in fts3.html works as expected. The tests run three
# times, with different values for DO_MALLOC_TEST.
# 
#   DO_MALLOC_TEST=0: Run tests with no OOM errors.
#   DO_MALLOC_TEST=1: Run tests with transient OOM errors.
#   DO_MALLOC_TEST=2: Run tests with persistent OOM errors.
#
foreach DO_MALLOC_TEST {0 1 2} {

# Reset the database and database connection. If this iteration of the 
# [foreach] loop is testing with OOM errors, disable the lookaside buffer.
#
db close
file delete -force test.db test.db-journal
sqlite3 db test.db
if {$DO_MALLOC_TEST} { sqlite3_db_config_lookaside db 0 0 0 }

##########################################################################
# Test the example CREATE VIRTUAL TABLE statements in section 1.1 
# of fts3.in.
#
ddl_test   1.1.1.1 {CREATE VIRTUAL TABLE data USING fts3()}
read_test  1.1.1.2 {PRAGMA table_info(data)} {0 content {} 0 {} 0}

ddl_test   1.1.2.1 {
  CREATE VIRTUAL TABLE pages USING fts3(title, keywords, body)
}
read_test  1.1.2.2 {
  PRAGMA table_info(pages)
} {0 title {} 0 {} 0 1 keywords {} 0 {} 0 2 body {} 0 {} 0}

ddl_test   1.1.3.1 {
  CREATE VIRTUAL TABLE mail USING fts3(
      subject VARCHAR(256) NOT NULL,
      body TEXT CHECK(length(body)<10240)
  )
}
read_test  1.1.3.2 {
  PRAGMA table_info(mail)
} {0 subject {} 0 {} 0 1 body {} 0 {} 0}

# A very large string. Used to test if the constraint on column "body" of
# table "mail" is enforced (it should not be - FTS3 tables do not support
# constraints).
set largetext [string repeat "the quick brown fox " 5000]
write_test 1.1.3.3 mail_content { INSERT INTO mail VALUES(NULL, $largetext) }
read_test  1.1.3.4 {
  SELECT subject IS NULL, length(body) FROM mail
} [list 1 100000]

ddl_test   1.1.4.1 {
  CREATE VIRTUAL TABLE papers USING fts3(author, document, tokenize=porter)
}
read_test  1.1.4.2 {
  PRAGMA table_info(papers)
} {0 author {} 0 {} 0 1 document {} 0 {} 0}

ddl_test   1.1.5.1 {
  CREATE VIRTUAL TABLE simpledata USING fts3(tokenize=simple)
}
read_test  1.1.5.2 {
  PRAGMA table_info(simpledata)
} {0 content {} 0 {} 0}

ifcapable icu {
  ddl_test 1.1.6.1 {
    CREATE VIRTUAL TABLE names USING fts3(a, b, tokenize=icu en_AU)
  }
  read_test  1.1.6.2 {
    PRAGMA table_info(names)
  } {0 a {} 0 {} 0 1 b {} 0 {} 0}
}

ddl_test   1.1.7.1 {DROP TABLE data}
ddl_test   1.1.7.2 {DROP TABLE pages}
ddl_test   1.1.7.3 {DROP TABLE mail}
ddl_test   1.1.7.4 {DROP TABLE papers}
ddl_test   1.1.7.5 {DROP TABLE simpledata}
read_test  1.1.7.6 {SELECT * FROM sqlite_master} {}

# The following is not one of the examples in section 1.1. It tests 
# specifying an FTS3 table with no module arguments using a slightly
# different syntax.
ddl_test   1.1.8.1 {CREATE VIRTUAL TABLE data USING fts3;}
read_test  1.1.8.2 {PRAGMA table_info(data)} {0 content {} 0 {} 0}
ddl_test   1.1.8.3 {DROP TABLE data}

##########################################################################

##########################################################################
# Test the examples in section 1.2 (populating fts3 tables)
#
ddl_test   1.2.1.1 {
  CREATE VIRTUAL TABLE pages USING fts3(title, body);
}
write_test 1.2.1.2 pages_content {
  INSERT INTO pages(docid, title, body) 
  VALUES(53, 'Home Page', 'SQLite is a software...');
}
read_test  1.2.1.3 {
  SELECT docid, * FROM pages
} {53 {Home Page} {SQLite is a software...}}

write_test 1.2.1.4 pages_content {
  INSERT INTO pages(title, body) 
  VALUES('Download', 'All SQLite source code...');
}
read_test  1.2.1.5 {
  SELECT docid, * FROM pages
} {53 {Home Page} {SQLite is a software...} 54 Download {All SQLite source code...}}

write_test 1.2.1.6 pages_content {
  UPDATE pages SET title = 'Download SQLite' WHERE rowid = 54
}
read_test  1.2.1.7 {
  SELECT docid, * FROM pages
} {53 {Home Page} {SQLite is a software...} 54 {Download SQLite} {All SQLite source code...}}

write_test 1.2.1.8 pages_content { DELETE FROM pages }
read_test  1.2.1.9 { SELECT docid, * FROM pages } {}

do_error_test fts3-1.2.1.10 {
  INSERT INTO pages(rowid, docid, title, body) VALUES(1, 2, 'A title', 'A document body');
} {SQL logic error or missing database}

# Test the optimize() function example:
if 0 {
ddl_test   1.2.2.1 { CREATE VIRTUAL TABLE docs USING fts3 }
write_test 1.2.2.2 docs_content {
  INSERT INTO docs VALUES('Others translate the first clause as');
}
write_test 1.2.2.3 docs_content {
  INSERT INTO docs VALUES('"which is for Solomon," meaning that');
}
write_test 1.2.2.4 docs_content {
  INSERT INTO docs VALUES('the book is dedicated to Solomon.');
}
read_test  1.2.2.5 { SELECT count(*) FROM docs_segdir } {3}
set DO_MALLOC_TEST 1
write_test 1.2.2.6 docs_segdir {
  INSERT INTO docs 
  SELECT * FROM (SELECT optimize(docs) FROM docs LIMIT 1) WHERE 0;
}
read_test  1.2.2.7 { SELECT count(*) FROM docs_segdir } {1}
}

##########################################################################
# Test the examples in section 1.3 (querying FTS3 tables)
#
ddl_test   1.3.1.1 { CREATE VIRTUAL TABLE mail USING fts3(subject, body) }
read_test  1.3.1.2 { 
  SELECT * FROM mail WHERE rowid = 15;                -- Fast. Rowid lookup.
  SELECT * FROM mail WHERE body MATCH 'sqlite';       -- Fast. Full-text query.
  SELECT * FROM mail WHERE mail MATCH 'search';       -- Fast. Full-text query.
  SELECT * FROM mail WHERE rowid BETWEEN 15 AND 20;   -- Slow. Linear scan.
  SELECT * FROM mail WHERE subject = 'database';      -- Slow. Linear scan.
  SELECT * FROM mail WHERE subject MATCH 'database';  -- Fast. Full-text query.
} {}
ddl_test   1.3.1.3 { DROP TABLE mail }

ddl_test   1.3.2.1 { CREATE VIRTUAL TABLE mail USING fts3(subject, body) }

write_test 1.3.2.2 mail_content {
  INSERT INTO mail(docid, subject, body) 
  VALUES(1, 'software feedback', 'found it too slow')
}
write_test 1.3.2.3 mail_content {
  INSERT INTO mail(docid, subject, body) 
  VALUES(2, 'software feedback', 'no feedback')
}
write_test 1.3.2.4 mail_content {
  INSERT INTO mail(docid, subject, body) 
  VALUES(3, 'slow lunch order',  'was a software problem')
}
read_test  1.3.2.5 {
  SELECT * FROM mail WHERE subject MATCH 'software'
} {{software feedback} {found it too slow} {software feedback} {no feedback}}
read_test  1.3.2.6 {
  SELECT * FROM mail WHERE body MATCH 'feedback'
} {{software feedback} {no feedback}}
read_test  1.3.2.7 {
  SELECT * FROM mail WHERE mail MATCH 'software'
} {{software feedback} {found it too slow} {software feedback} {no feedback} {slow lunch order} {was a software problem}}
read_test  1.3.2.7 {
  SELECT * FROM mail WHERE mail MATCH 'slow'
} {{software feedback} {found it too slow} {slow lunch order} {was a software problem}}
ddl_test   1.3.2.8 { DROP TABLE mail }

ddl_test   1.3.3.1 { CREATE VIRTUAL TABLE docs USING fts3(content) }
read_test  1.3.3.2 { SELECT * FROM docs WHERE docs MATCH 'sqlite' } {}
read_test  1.3.3.3 { SELECT * FROM docs WHERE docs.docs MATCH 'sqlite' } {}
read_test  1.3.3.4 { SELECT * FROM docs WHERE main.docs.docs MATCH 'sqlite' } {}
do_error_test e_fts3-1.3.3.5 { 
  SELECT * FROM docs WHERE main.docs MATCH 'sqlite' 
} {no such column: main.docs}
ddl_test   1.3.2.8 { DROP TABLE docs }

##########################################################################

##########################################################################
# Test the examples in section 3 (full-text index queries).
#
ddl_test   1.4.1.1 { CREATE VIRTUAL TABLE docs USING fts3(title, body) }
foreach {tn title body} {
  2 "linux driver" "a device"
  3 "driver"       "linguistic trick"
  4 "problems"     "linux problems"
  5 "linux"        "big problems"
  6 "linux driver" "a device driver problem"
  7 "good times"   "applications for linux"
  8 "not so good"  "linux applications"
  9 "alternative"  "linoleum appliances"
 10 "no L I N"     "to be seen"
} {
  write_test 1.4.1.$tn docs_content { INSERT INTO docs VALUES($title,$body) }
  set R($tn) [list $title $body]
}

read_test  1.4.1.11 { 
  SELECT * FROM docs WHERE docs MATCH 'linux'
} [concat $R(2) $R(4) $R(5) $R(6) $R(7) $R(8)]
read_test  1.4.1.12 { 
  SELECT * FROM docs WHERE docs MATCH 'lin*'
} [concat $R(2) $R(3) $R(4) $R(5) $R(6) $R(7) $R(8) $R(9)]
read_test  1.4.1.13 { 
  SELECT * FROM docs WHERE docs MATCH 'title:linux problems'
} [concat $R(5)]
read_test  1.4.1.14 { 
  SELECT * FROM docs WHERE body MATCH 'title:linux driver'
} [concat $R(6)]
read_test  1.4.1.15 { 
  SELECT * FROM docs WHERE docs MATCH '"linux applications"'
} [concat $R(8)]
read_test  1.4.1.16 { 
  SELECT * FROM docs WHERE docs MATCH '"lin* app*"'
} [concat $R(8) $R(9)]
ddl_test   1.4.1.17 { DROP TABLE docs }

ddl_test   1.4.2.1 { CREATE VIRTUAL TABLE docs USING fts3() }
write_test 1.4.2.2 docs_content { 
  INSERT INTO docs VALUES(
  'SQLite is an ACID compliant embedded relational database management system')
}
foreach {tn query hit} {
3 {SELECT * FROM docs WHERE docs MATCH 'sqlite NEAR database'} 1
4 {SELECT * FROM docs WHERE docs MATCH 'database NEAR/6 sqlite'} 1
5 {SELECT * FROM docs WHERE docs MATCH 'database NEAR/5 sqlite'} 0
6 {SELECT * FROM docs WHERE docs MATCH 'database NEAR/2 "ACID compliant"'} 1
7 {SELECT * FROM docs WHERE docs MATCH '"ACID compliant" NEAR/2 sqlite'} 1
8 {SELECT * FROM docs WHERE docs MATCH 'sqlite NEAR/2 acid NEAR/2 relational'} 1
9 {SELECT * FROM docs WHERE docs MATCH 'acid NEAR/2 sqlite NEAR/2 relational'} 0
} {
  set res [db eval {SELECT * FROM docs WHERE $hit}]
  read_test 1.4.2.$tn $query $res
}
ddl_test 1.4.2.10 { DROP TABLE docs }

unset R


# TODO: Change numbering after here...

##########################################################################
# Test the example in section 5 (custom tokenizers).
#
ddl_test   2.1.1 { CREATE VIRTUAL TABLE simple USING fts3(tokenize=simple) } 
write_test 2.1.2 simple_content { 
  INSERT INTO simple VALUES('Right now they''re very frustrated')
}
read_test  2.1.3 {SELECT docid FROM simple WHERE simple MATCH 'Frustrated'} {1} 
read_test  2.1.4 {SELECT docid FROM simple WHERE simple MATCH 'Frustration'} {}

ddl_test   2.2.1 { CREATE VIRTUAL TABLE porter USING fts3(tokenize=porter) } 
write_test 2.2.2 porter_content { 
  INSERT INTO porter VALUES('Right now they''re very frustrated')
}
read_test  2.2.3 {SELECT docid FROM porter WHERE porter MATCH 'Frustrated'} {1} 
read_test  2.2.4 {SELECT docid FROM porter WHERE porter MATCH 'Frustration'} {1}
##########################################################################

##########################################################################
# Test the examples in section 4 (auxillary functions).
#
ddl_test   3.1.1 { CREATE VIRTUAL TABLE mail USING fts3(subject, body) }

write_test 3.1.2 mail_content { 

read_test 3.2.4 {
  SELECT snippet(text, '[', ']', '...') FROM text WHERE text MATCH '"min* tem*"'
} {{... 2-3oC drops. Cool in the upper portion, [minimum] [temperature] 14-16oC and cool elsewhere, [minimum] ...}}

#break
}
# End of tests of example code in fts3.html
#-------------------------------------------------------------------------


finish_test