test032.tcl
上传用户:tsgydb
上传日期:2007-04-14
资源大小:10674k
文件大小:6k
- # See the file LICENSE for redistribution information.
- #
- # Copyright (c) 1996, 1997, 1998, 1999, 2000
- # Sleepycat Software. All rights reserved.
- #
- # $Id: test032.tcl,v 11.15 2000/08/25 14:21:55 sue Exp $
- #
- # DB Test 32 {access method}
- # Use the first 10,000 entries from the dictionary.
- # Insert each with self as key and "ndups" duplicates
- # For the data field, prepend the letters of the alphabet
- # in a random order so that we force the duplicate sorting
- # code to do something.
- # By setting ndups large, we can make this an off-page test
- # After all are entered; test the DB_GET_BOTH functionality
- # first by retrieving each dup in the file explicitly. Then
- # test the failure case.
- proc test032 { method {nentries 10000} {ndups 5} {tnum 32} args } {
- global alphabet rand_init
- source ./include.tcl
- set args [convert_args $method $args]
- set omethod [convert_method $method]
- berkdb srand $rand_init
- # Create the database and open the dictionary
- set eindex [lsearch -exact $args "-env"]
- #
- # If we are using an env, then testfile should just be the db name.
- # Otherwise it is the test directory and the name.
- if { $eindex == -1 } {
- set testfile $testdir/test0$tnum.db
- set checkdb $testdir/checkdb.db
- set env NULL
- } else {
- set testfile test0$tnum.db
- set checkdb checkdb.db
- incr eindex
- set env [lindex $args $eindex]
- }
- set t1 $testdir/t1
- set t2 $testdir/t2
- set t3 $testdir/t3
- cleanup $testdir $env
- puts "Test0$tnum:
- $method ($args) $nentries small sorted dup key/data pairs"
- if { [is_record_based $method] == 1 ||
- [is_rbtree $method] == 1 } {
- puts "Test0$tnum skipping for method $omethod"
- return
- }
- set db [eval {berkdb_open -create -truncate -mode 0644
- $omethod -dup -dupsort} $args {$testfile} ]
- error_check_good dbopen [is_valid_db $db] TRUE
- set did [open $dict]
- set check_db [eval {berkdb_open
- -create -truncate -mode 0644} $args {-hash $checkdb}]
- error_check_good dbopen:check_db [is_valid_db $check_db] TRUE
- set pflags ""
- set gflags ""
- set txn ""
- set count 0
- # Here is the loop where we put and get each key/data pair
- puts "tTest0$tnum.a: Put/get loop"
- set dbc [eval {$db cursor} $txn]
- error_check_good cursor_open [is_substr $dbc $db] 1
- while { [gets $did str] != -1 && $count < $nentries } {
- # Re-initialize random string generator
- randstring_init $ndups
- set dups ""
- for { set i 1 } { $i <= $ndups } { incr i } {
- set pref [randstring]
- set dups $dups$pref
- set datastr $pref:$str
- set ret [eval {$db put}
- $txn $pflags {$str [chop_data $method $datastr]}]
- error_check_good put $ret 0
- }
- set ret [eval {$check_db put}
- $txn $pflags {$str [chop_data $method $dups]}]
- error_check_good checkdb_put $ret 0
- # Now retrieve all the keys matching this key
- set x 0
- set lastdup ""
- for {set ret [$dbc get -set $str]}
- {[llength $ret] != 0}
- {set ret [$dbc get -nextdup] } {
- set k [lindex [lindex $ret 0] 0]
- if { [string compare $k $str] != 0 } {
- break
- }
- set datastr [lindex [lindex $ret 0] 1]
- if {[string length $datastr] == 0} {
- break
- }
- if {[string compare $lastdup $datastr] > 0} {
- error_check_good sorted_dups($lastdup,$datastr)
- 0 1
- }
- incr x
- set lastdup $datastr
- }
- error_check_good "Test0$tnum:ndups:$str" $x $ndups
- incr count
- }
- error_check_good cursor_close [$dbc close] 0
- close $did
- # Now we will get each key from the DB and compare the results
- # to the original.
- puts "tTest0$tnum.b: Checking file for correct duplicates (no cursor)"
- set check_c [eval {$check_db cursor} $txn]
- error_check_good check_c_open(2)
- [is_substr $check_c $check_db] 1
- for {set ndx 0} {$ndx < [expr 4 * $ndups]} {incr ndx 4} {
- for {set ret [$check_c get -first]}
- {[llength $ret] != 0}
- {set ret [$check_c get -next] } {
- set k [lindex [lindex $ret 0] 0]
- set d [lindex [lindex $ret 0] 1]
- error_check_bad data_check:$d [string length $d] 0
- set pref [string range $d $ndx [expr $ndx + 3]]
- set data $pref:$k
- set ret [eval {$db get} $txn {-get_both $k $data}]
- error_check_good
- get_both_data:$k $ret [list [list $k $data]]
- }
- }
- $db sync
- # Now repeat the above test using cursor ops
- puts "tTest0$tnum.c: Checking file for correct duplicates (cursor)"
- set dbc [eval {$db cursor} $txn]
- error_check_good cursor_open [is_substr $dbc $db] 1
- for {set ndx 0} {$ndx < [expr 4 * $ndups]} {incr ndx 4} {
- for {set ret [$check_c get -first]}
- {[llength $ret] != 0}
- {set ret [$check_c get -next] } {
- set k [lindex [lindex $ret 0] 0]
- set d [lindex [lindex $ret 0] 1]
- error_check_bad data_check:$d [string length $d] 0
- set pref [string range $d $ndx [expr $ndx + 3]]
- set data $pref:$k
- set ret [eval {$dbc get} {-get_both $k $data}]
- error_check_good
- get_both_key:$k $ret [list [list $k $data]]
- }
- }
- # Now check the error case
- puts "tTest0$tnum.d: Check error case (no cursor)"
- for {set ret [$check_c get -first]}
- {[llength $ret] != 0}
- {set ret [$check_c get -next] } {
- set k [lindex [lindex $ret 0] 0]
- set d [lindex [lindex $ret 0] 1]
- error_check_bad data_check:$d [string length $d] 0
- set data XXX$k
- set ret [eval {$db get} $txn {-get_both $k $data}]
- error_check_good error_case:$k [llength $ret] 0
- }
- # Now check the error case
- puts "tTest0$tnum.e: Check error case (cursor)"
- for {set ret [$check_c get -first]}
- {[llength $ret] != 0}
- {set ret [$check_c get -next] } {
- set k [lindex [lindex $ret 0] 0]
- set d [lindex [lindex $ret 0] 1]
- error_check_bad data_check:$d [string length $d] 0
- set data XXX$k
- set ret [eval {$dbc get} {-get_both $k $data}]
- error_check_good error_case:$k [llength $ret] 0
- }
- error_check_good check_c:close [$check_c close] 0
- error_check_good check_db:close [$check_db close] 0
- error_check_good dbc_close [$dbc close] 0
- error_check_good db_close [$db close] 0
- }