From 3d952cd8bd2d522c9334f850b88d0bc16155f760 Mon Sep 17 00:00:00 2001
From: Michael Okoko <okokomichaels@outlook.com>
Date: Thu, 19 Aug 2021 14:38:11 +0100
Subject: [PATCH] Improve tests and test results to cover larger cases

Signed-off-by: Michael Okoko <okokomichaels@outlook.com>
---
 mysql-test/main/statistics_fetch.result | Bin 3325 -> 4545 bytes
 mysql-test/main/statistics_fetch.test   |  57 +-
 mysql-test/main/statistics_json.result  | 972 ++++++++++++++++++------
 mysql-test/main/statistics_json.test    |  90 ++-
 sql/sql_statistics.cc                   |   5 +-
 5 files changed, 834 insertions(+), 290 deletions(-)

diff --git a/mysql-test/main/statistics_fetch.result b/mysql-test/main/statistics_fetch.result
index 4e835b6a35f4fef5439ad79af8131800bf6eed46..2c79ab4456161fbfcef9790a3ea1bf55838bbe7e 100644
GIT binary patch
literal 4545
zcmc&&+iu%95Phb;f=C|h6p0~OmM=95B-k~uZPTPcTJ&MDP#BB0wJT8yNjrAFzQc=r
zw;Vgj#sI}B#Tju9c@8H%bx!dQDkPQ}j+w}5AIuTOoUkDc1x;Wep2Tv7IhAA1XoMeT
z_^(*VY{*F({wp$u35%A79%4rL4C2Zx%06N-%Ek$JQJhjf1plX4j<AqKLXfyGbcmXS
zH8`h^pb}pSGNIvU{s}qD>8H!f)q(ETsUm(}ovH@9PM=dEDVF3RfhT2E0=Y_@^Eic-
z;{WLEtm{DbluPv@Q*-cyBx5RY)jd0}deG|7d}s}50kkHx7PL0B4*0Kr#0voi3?G^h
z-X9TOtp~x``d7>(12XPstWRXMLM|%jn4IAO&r-Dzlj$ORoulV=9l}U5d!DCdg5{J8
zK#oFsgN$R6Bwgo+${XG0=$jaZ3Qc3ssdN<VHK|>Zjfd3Mj0NhkVa(`7%p#goctFI(
zron0xbu?&H7UK3R?2?*>M|4`9r(knSXW}ubL7;KULKRxzrxE2gp^%Z3eyW(YRl3e<
z1gGVBe+2OJ+wI;J{Mvu^e1}f+gaGW2nObp#?bkn8FbzcJBLq&Qe4?<?fVfX8`Vd#1
ziywS3j*#&;nCa7lG<<m`L=2dQaB>8KhiYL$8RB$+c=jaFAWkIY5~1Eya^iUC?(?_r
z*F9vhzPi3fH$OKYo{Ivv=DO&oh}qEK;rt?;!N0^u=O&XBd9Lf*mEn7}f%mn06BZ0|
zkFV~q>h$?pdv5_=8`WOrXv`1W`ytB1RbY8|JjR~J*dX29z<Zs^TI2RoS<kZYyp+pY
z+LYrKHtTpt(pmYwme1Q_EC_8W^->P2U_hzpLDL#(E*aDEU$GsfF$=ZqqLfUn#l}e@
zLJqe@hR9^7ST>X+Z4Nkn98=bxAtf!L5=LWblFsEuE0iLh$q11P7g`6YeFvubrt@s)
zqAz#1H|TE%;}e*9a&8Q-FzSWj7e=Eng2HGPMyoK|h0!^2{&|5;sij)7zMoKOyh@a+
z*{-#`Ba}+TEtM)gVN<IXe#eTHitzuydXQSFWuX$Fja(VH_sLaV6&|L*SuZ#AtKv>1
zA3S$wjiGgETOGEcQzqt7GNoN#r&3!(fJp5t1k9>83V}w$mH|QVX7|5Ya>-`<e$Dp}
zW|a<GLMroqol{F*b4o3Pu<P2_ew##Ec70|2bO3+!V)oO=;A*I}TwSf+_H}(r?Bb8s
z4|~V&Cy<)Gg)lnEk&Mww@v1`IB)CqyBQJIn?42+6A|vFYj$7R7jor7pz0*D0*KP0H
zzG3^o_D$QjY~Qwh$9>6FZWi~ZRM}YuTVP7G##2?#5n7{;VX4^Y37cxO@H<d$l;fL%
oc#z(x)$A7>1Ma5aL_Euj|BKLnkwRU(GNi(<j%kToA2QQ^19{iqod5s;

delta 760
zcmX@8{8w^9|HP*ylVcebC;PD~2_)tv=2TXtDwHH9<p4>;$>$l(CpU1}F&a*mWAm7t
z&DJpa1Dis<k*Nu%i+`w-k86BDkgIdNhZ83Q4p1nA3qH2M1w(OyV}&4C5227ri#}X%
z6;2?192e|;j0^q;0bwwSN=&xrR|JOAZ1#r93LG-5rj~lf7Ly%0j3%FDmJv!&D9=bO
zN>xZGNi8l(&`SUY>;?{1wdA7I#1eR58)+meWagD<T61OQ6{i-JfH?UIB}S7!a*Crz
zj1e$m%ovS;QDZ;(Fqg^X-(2NjYYh$bEKPvc8Up3ca?6NgSZoBeJuzW&9>16#S4vTS
z0X(LyQMg9dT&RvSgeDPhtSMp{)Zf8W@vYY67+x{#9%M2!n<QXL$ZPT_elf)G8UO)_
BB%1&L

diff --git a/mysql-test/main/statistics_fetch.test b/mysql-test/main/statistics_fetch.test
index 4a57906f95a..d0313fe5f2b 100644
--- a/mysql-test/main/statistics_fetch.test
+++ b/mysql-test/main/statistics_fetch.test
@@ -3,38 +3,47 @@
 --echo # todo: should be merged with statistics_json.test
 --echo #
 
+set @save_histogram_type=@@histogram_type;
+set @save_histogram_size=@@histogram_size;
+
 create table ten(a int primary key);
 insert into ten values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
 
-create table t1 (a varchar(255));
-insert into t1 select concat('a-', a) from ten;
+create table t1_bin (a varchar(255));
+insert into t1_bin select concat('a-', a) from ten;
 
-analyze table t1 persistent for all;
-select * from mysql.column_stats where table_name='t1';
-explain extended select * from t1 where a between 'a-3a' and 'zzzzzzzzz';
+set histogram_size=10;
+analyze table t1_bin persistent for all;
+select hex(histogram) from mysql.column_stats where table_name='t1_bin';
+explain extended select * from t1_bin where a between 'a-3a' and 'zzzzzzzzz';
+analyze select * from t1_bin where a between 'a-3a' and 'zzzzzzzzz';
 
-create table t2(a int);
-insert into t2 select a*10 from ten;
-analyze table t2 persistent for all;
-explain extended select * from t2 where a between '44' and '55';
-
-drop table t1;
-drop table t2;
-
-create table t1 (a varchar(255));
-insert into t1 select concat('a-', a) from ten;
+create table t1_json (a varchar(255));
+insert into t1_json select concat('a-', a) from ten;
 set histogram_type=json;
+analyze table t1_json persistent for all;
+select * from mysql.column_stats where table_name='t1_json';
+explain extended select * from t1_json where a between 'a-3a' and 'zzzzzzzzz';
+analyze select * from t1_json where a between 'a-3a' and 'zzzzzzzzz';
 
-analyze table t1 persistent for all;
-select * from mysql.column_stats where table_name='t1';
-explain extended select * from t1 where a between 'a-3a' and 'zzzzzzzzz';
 
-create table t2(a int);
-insert into t2 select a*10 from ten;
+create table t2_bin(a int);
+insert into t2_bin select a*10 from ten;
+set histogram_type=@save_histogram_type;
+analyze table t2_bin persistent for all;
+explain extended select * from t2_bin where a between '44' and '55';
+analyze select * from t2_bin where a between '44' and '55';
+
+create table t2_json(a int);
+insert into t2_json select a*10 from ten;
 set histogram_type=json;
-analyze table t2 persistent for all;
-explain extended select * from t2 where a between '44' and '55';
+analyze table t2_json persistent for all;
+select * from mysql.column_stats where table_name='t2_json';
+explain extended select * from t2_json where a between '44' and '55';
+analyze select * from t2_json where a between '44' and '55';
 
-drop table t1;
-drop table t2;
+drop table t1_bin;
+drop table t1_json;
+drop table t2_bin;
+drop table t2_json;
 
diff --git a/mysql-test/main/statistics_json.result b/mysql-test/main/statistics_json.result
index 0f68398650d..6a8df8f0f76 100644
--- a/mysql-test/main/statistics_json.result
+++ b/mysql-test/main/statistics_json.result
@@ -5,309 +5,793 @@ drop table if exists t1;
 set @save_histogram_type=@@histogram_type;
 set @save_histogram_size=@@histogram_size;
 CREATE TABLE t1 (
-a int,
+a int NOT NULL PRIMARY KEY,
 b varchar(32),
-c char(2),
-d double
+c char(16),
+d date,
+e double,
+f bit(3),
+INDEX idx1 (b, e),
+INDEX idx2 (c, d),
+INDEX idx3 (d),
+INDEX idx4 (e, b, d)
 );
-INSERT INTO t1 SELECT seq, seq, seq, seq from seq_1_to_25;
+INSERT INTO t1 VALUES
+(0, NULL, NULL, NULL, NULL, NULL),
+(7, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'dddddddd', '1990-05-15', 0.1, b'100'),
+(17, 'vvvvvvvvvvvvv', 'aaaa', '1989-03-12', 0.01, b'101'),
+(1, 'vvvvvvvvvvvvv', NULL, '1989-03-12', 0.01, b'100'),
+(12, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1999-07-23', 0.112, b'001'),
+(23, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.1, b'100'),
+(8, 'vvvvvvvvvvvvv', 'aaaa', '1999-07-23', 0.1, b'100'),
+(22, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'aaaa', '1989-03-12', 0.112, b'001'),
+(31, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'aaaa', '1999-07-23', 0.01, b'001'),
+(10, NULL, 'aaaa', NULL, 0.01, b'010'),
+(5, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1999-07-23', 0.1, b'100'),
+(15, 'vvvvvvvvvvvvv', 'ccccccccc', '1990-05-15', 0.1, b'010'),
+(30, NULL, 'bbbbbb', NULL, NULL, b'100'),
+(38, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', NULL, NULL, NULL),
+(18, 'zzzzzzzzzzzzzzzzzz', 'ccccccccc', '1990-05-15', 0.01, b'010'),
+(9, 'yyy', 'bbbbbb', '1998-08-28', 0.01, NULL),
+(29, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.012, b'010'),
+(3, 'yyy', 'dddddddd',  '1990-05-15', 0.112, b'010'),
+(39, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', NULL, 0.01, b'100'),
+(14, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'ccccccccc', '1990-05-15', 0.1, b'100'),
+(40, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', '1989-03-12', NULL, NULL),
+(44, NULL, 'aaaa', '1989-03-12', NULL, b'010'),
+(19, 'vvvvvvvvvvvvv', 'ccccccccc', '1990-05-15', 0.012, b'011'),
+(21, 'zzzzzzzzzzzzzzzzzz', 'dddddddd', '1989-03-12', 0.112, b'100'),
+(45, NULL, NULL, '1989-03-12', NULL, b'011'),
+(2, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'ccccccccc', '1990-05-15', 0.1, b'001'),
+(35, 'yyy', 'aaaa', '1990-05-15', 0.05, b'011'),
+(4, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.01, b'101'),
+(47, NULL, 'aaaa', '1990-05-15', 0.05, b'010'),
+(42, NULL, 'ccccccccc', '1989-03-12', 0.01, b'010'),
+(32, NULL, 'bbbbbb', '1990-05-15', 0.01, b'011'),
+(49, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww' , 'aaaa', '1990-05-15', NULL, NULL),
+(43, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww' , 'bbbbbb', '1990-05-15', NULL, b'100'),
+(37, 'yyy', NULL, '1989-03-12', 0.05, b'011'),
+(41, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'ccccccccc', '1990-05-15', 0.05, NULL),
+(34, 'yyy', NULL, NULL, NULL, NULL),
+(33, 'zzzzzzzzzzzzzzzzzz', 'dddddddd', '1989-03-12', 0.05, b'011'),
+(24,  'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1990-05-15', 0.01,  b'101'),
+(11, 'yyy', 'ccccccccc', '1999-07-23', 0.1, NULL),
+(25, 'zzzzzzzzzzzzzzzzzz', 'bbb', '1989-03-12', 0.01,  b'101');
 SET histogram_type='JSON';
-set histogram_size=10;
-ANALYZE TABLE t1 PERSISTENT FOR ALL;
+set histogram_size=25;
+ANALYZE TABLE t1 persistent for all;
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	Engine-independent statistics collected
 test.t1	analyze	status	OK
-SELECT * FROM mysql.column_stats WHERE table_name='t1';
+SELECT * FROM mysql.table_stats;
+db_name	table_name	cardinality
+test	t1	40
+SELECT * FROM mysql.column_stats;
 db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
-test	t1	a	1	25	0.0000	4.0000	1.0000	10	JSON	[
+test	t1	a	0	49	0.0000	4.0000	1.0000	25	JSON	[
+  "1",
   "3",
-  "5",
+  "4",
   "7",
+  "8",
   "10",
-  "12",
-  "14",
-  "16",
-  "19",
-  "21",
-  "23"
-]
-test	t1	b	1	9	0.0000	1.6400	1.0000	10	JSON	[
   "11",
-  "13",
+  "14",
   "15",
   "18",
-  "2",
-  "21",
-  "23",
-  "3",
-  "5",
-  "7"
-]
-test	t1	c	1	9	0.0000	2.0000	1.0000	10	JSON	[
-  "11",
-  "13",
-  "15",
-  "18",
-  "2",
-  "21",
-  "23",
-  "3",
-  "5",
-  "7"
-]
-test	t1	d	1	25	0.0000	8.0000	1.0000	10	JSON	[
-  "3",
-  "5",
-  "7",
-  "10",
-  "12",
-  "14",
-  "16",
   "19",
-  "21",
-  "23"
+  "22",
+  "24",
+  "25",
+  "30",
+  "31",
+  "33",
+  "34",
+  "37",
+  "38",
+  "40",
+  "41",
+  "43",
+  "44",
+  "47"
 ]
-explain extended select * from t1 where b between '20' and '70';
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	5	JSON	[
+  "vvvvvvvvvvvvv",
+  "wwwwwwwwwwwwwwwwwwwwwwwwwwww",
+  "xxxxxxxxxxxxxxxxxxxxxxxxxx",
+  "yyy",
+  "zzzzzzzzzzzzzzzzzz"
+]
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	5	JSON	[
+  "aaaa",
+  "bbb",
+  "bbbbbb",
+  "ccccccccc",
+  "dddddddd"
+]
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	4	JSON	[
+  "1989-03-12",
+  "1990-05-15",
+  "1998-08-28",
+  "1999-07-23"
+]
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	5	JSON	[
+  "0.01",
+  "0.012",
+  "0.05",
+  "0.1",
+  "0.112"
+]
+test	t1	f	1	5	0.2000	1.0000	6.4000	5	JSON	[
+  "
",
+  "",
+  "",
+  "",
+  ""
+]
+SELECT * FROM mysql.index_stats;
+db_name	table_name	index_name	prefix_arity	avg_frequency
+test	t1	PRIMARY	1	1.0000
+test	t1	idx1	1	6.4000
+test	t1	idx1	2	1.6875
+test	t1	idx2	1	7.0000
+test	t1	idx2	2	2.3846
+test	t1	idx3	1	8.5000
+test	t1	idx4	1	6.2000
+test	t1	idx4	2	1.6875
+test	t1	idx4	3	1.1304
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+40
+explain extended select * from t1 where a between '20' and '70';
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	filtered	Extra
-1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	25	10.00	Using where
+1	SIMPLE	t1	ALL	PRIMARY	NULL	NULL	NULL	40	57.50	Using where
 Warnings:
-Note	1003	select `test`.`t1`.`a` AS `a`,`test`.`t1`.`b` AS `b`,`test`.`t1`.`c` AS `c`,`test`.`t1`.`d` AS `d` from `test`.`t1` where `test`.`t1`.`b` between '20' and '70'
-UPDATE mysql.column_stats SET histogram='["1", {"a": "b"}, "2"]' WHERE table_name='t1';
-FLUSH TABLES;
-SELECT * FROM t1;
-ERROR HY000: Failed to parse histogram, encountered JSON_TYPE '1'.
-UPDATE mysql.column_stats SET histogram='{}' WHERE table_name='t1';
-FLUSH TABLES;
-SELECT * FROM t1;
-ERROR HY000: Failed to parse histogram, encountered JSON_TYPE '1'.
+Note	1003	select `test`.`t1`.`a` AS `a`,`test`.`t1`.`b` AS `b`,`test`.`t1`.`c` AS `c`,`test`.`t1`.`d` AS `d`,`test`.`t1`.`e` AS `e`,`test`.`t1`.`f` AS `f` from `test`.`t1` where `test`.`t1`.`a` between '20' and '70'
+analyze select * from t1 where a between '20' and '70';
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	r_rows	filtered	r_filtered	Extra
+1	SIMPLE	t1	ALL	PRIMARY	NULL	NULL	NULL	40	40.00	57.50	57.50	Using where
 DELETE FROM mysql.column_stats;
 DROP TABLE t1;
 create schema world;
 use world;
 set histogram_type='JSON';
-set histogram_size=25;
-ANALYZE TABLE Country PERSISTENT FOR ALL;
+set histogram_size=50;
+ANALYZE TABLE Country, City, CountryLanguage persistent for all;
 SELECT column_name, min_value, max_value, hist_size, hist_type, histogram FROM mysql.column_stats;
 column_name	min_value	max_value	hist_size	hist_type	histogram
-Code	ABW	ZWE	25	JSON	[
+Code	ABW	ZWE	50	JSON	[
+  "ALB",
   "ARM",
+  "AUS",
   "BEL",
-  "BLZ",
-  "CAF",
-  "COG",
-  "CZE",
-  "ERI",
-  "FRO",
-  "GMB",
-  "GUY",
-  "IRL",
-  "KAZ",
-  "LBN",
-  "LVA",
-  "MKD",
-  "MUS",
-  "NIC",
-  "PAN",
-  "PRT",
-  "SAU",
-  "SMR",
-  "SYC",
-  "TMP",
+  "BHR",
+  "BMU",
+  "BRN",
+  "CAN",
+  "CIV",
+  "COK",
+  "CUB",
+  "DEU",
+  "DOM",
+  "ESH",
+  "FJI",
+  "FSM",
+  "GIB",
+  "GNQ",
+  "GUF",
+  "HMD",
+  "IDN",
+  "IRQ",
+  "JAM",
+  "KGZ",
+  "KWT",
+  "LBY",
+  "LTU",
+  "MCO",
+  "MEX",
+  "MMR",
+  "MSR",
+  "MYS",
+  "NFK",
+  "NOR",
+  "PAK",
+  "PHL",
+  "PRK",
+  "QAT",
+  "RWA",
+  "SGS",
+  "SLV",
+  "STP",
+  "SWZ",
+  "TGO",
+  "TKM",
+  "TUR",
   "UKR",
-  "VIR"
+  "UZB",
+  "VIR",
+  "YEM"
 ]
-Name	Afghanistan	Zimbabwe	25	JSON	[
+Name	Afghanistan	Zimbabwe	50	JSON	[
+  "Andorra",
   "Argentina",
+  "Azerbaijan",
   "Barbados",
-  "Botswana",
-  "Cameroon",
-  "Cocos (Keeling) Islands",
-  "Czech Republic",
-  "El Salvador",
-  "France",
-  "Gibraltar",
-  "Guyana",
-  "Iran",
-  "Kenya",
-  "Libyan Arab Jamahiriya",
-  "Maldives",
-  "Micronesia, Federated States of",
-  "Nepal",
-  "Norfolk Island",
-  "Papua New Guinea",
-  "Romania",
-  "Samoa",
-  "Solomon Islands",
-  "Svalbard and Jan Mayen",
-  "Togo",
+  "Bermuda",
+  "Bouvet Island",
+  "Bulgaria",
+  "Canada",
+  "Chile",
+  "Colombia",
+  "Costa Rica",
+  "Côte d’Ivoire",
+  "Dominican Republic",
+  "Equatorial Guinea",
+  "Faroe Islands",
+  "French Guiana",
+  "Georgia",
+  "Greenland",
+  "Guinea",
+  "Heard Island and McDonald Islands",
+  "Iceland",
+  "Ireland",
+  "Japan",
+  "Kuwait",
+  "Lesotho",
+  "Lithuania",
+  "Malawi",
+  "Marshall Islands",
+  "Mayotte",
+  "Mongolia",
+  "Namibia",
+  "Netherlands Antilles",
+  "Nigeria",
+  "Norway",
+  "Panama",
+  "Philippines",
+  "Qatar",
+  "Saint Helena",
+  "Saint Vincent and the Grenadines",
+  "Senegal",
+  "Slovenia",
+  "South Georgia and the South Sandwich Islands",
+  "Suriname",
+  "Syria",
+  "Thailand",
+  "Tunisia",
   "Uganda",
-  "Venezuela"
+  "United States",
+  "Venezuela",
+  "Western Sahara"
 ]
-SurfaceArea	0.40	17075400.00	25	JSON	[
+SurfaceArea	0.40	17075400.00	50	JSON	[
+  "14.00",
   "36.00",
+  "78.00",
   "151.00",
-  "261.00",
-  "388.00",
-  "618.00",
-  "1102.00",
-  "4000.00",
-  "10990.00",
-  "18274.00",
-  "27750.00",
-  "36188.00",
-  "51197.00",
-  "71740.00",
-  "93030.00",
-  "112622.00",
-  "163610.00",
-  "238391.00",
-  "283561.00",
-  "338145.00",
-  "449964.00",
-  "587041.00",
-  "796095.00",
-  "1104300.00",
+  "200.00",
+  "264.00",
+  "344.00",
+  "430.00",
+  "464.00",
+  "622.00",
+  "751.00",
+  "1399.00",
+  "2510.00",
+  "4033.00",
+  "8875.00",
+  "11000.00",
+  "14874.00",
+  "20256.00",
+  "25713.00",
+  "28051.00",
+  "30518.00",
+  "41526.00",
+  "48511.00",
+  "56785.00",
+  "69700.00",
+  "78866.00",
+  "90000.00",
+  "103000.00",
+  "111369.00",
+  "120538.00",
+  "147181.00",
+  "181035.00",
+  "214969.00",
+  "242900.00",
+  "274000.00",
+  "309500.00",
+  "331689.00",
+  "390757.00",
+  "447400.00",
+  "505992.00",
+  "581730.00",
+  "637657.00",
+  "774815.00",
+  "912050.00",
+  "1098581.00",
+  "1246700.00",
   "1648195.00",
-  "2724900.00"
+  "2149690.00",
+  "2724900.00",
+  "9363520.00"
 ]
-Population	0	1277558000	25	JSON	[
+Population	0	1277558000	50	JSON	[
+  "0",
   "1000",
+  "2500",
   "8000",
-  "25000",
-  "65000",
-  "94000",
-  "181000",
-  "293000",
-  "453000",
-  "817000",
-  "1439200",
-  "2542000",
-  "3337000",
-  "3869000",
-  "4807000",
-  "5496000",
-  "7430000",
-  "8861400",
-  "10278100",
-  "11669000",
-  "16125000",
-  "22256000",
-  "28351000",
-  "40377000",
+  "17000",
+  "27000",
+  "38000",
+  "68000",
+  "78000",
+  "99000",
+  "149000",
+  "190000",
+  "241000",
+  "307000",
+  "428000",
+  "456000",
+  "638000",
+  "885000",
+  "1295000",
+  "1726000",
+  "2153000",
+  "2662000",
+  "3101000",
+  "3520000",
+  "3850000",
+  "4023000",
+  "4629000",
+  "5074000",
+  "5398700",
+  "6188000",
+  "6782000",
+  "7733000",
+  "8329000",
+  "9586000",
+  "10239000",
+  "10730000",
+  "11385000",
+  "14786000",
+  "15942000",
+  "18886000",
+  "22244000",
+  "23115000",
+  "25662000",
+  "31471000",
+  "39441700",
+  "50456000",
   "61399000",
-  "111506000"
+  "68470000",
+  "111506000",
+  "170115000"
 ]
-Capital	1	4074	25	JSON	[
-  "62",
-  "149",
-  "201",
-  "553",
-  "645",
-  "902",
-  "922",
-  "1109",
-  "1530",
-  "1859",
-  "2296",
-  "2430",
-  "2452",
-  "2482",
-  "2689",
-  "2729",
-  "2831",
-  "2919",
-  "3047",
-  "3162",
-  "3209",
-  "3250",
-  "3336",
-  "3426",
-  "3538"
+Capital	1	4074	50	JSON	[
+  "35",
+  "63",
+  "129",
+  "150",
+  "187",
+  "204",
+  "538",
+  "554",
+  "586",
+  "652",
+  "764",
+  "904",
+  "917",
+  "926",
+  "937",
+  "1365",
+  "1464",
+  "1532",
+  "1800",
+  "1864",
+  "2256",
+  "2317",
+  "2409",
+  "2434",
+  "2441",
+  "2454",
+  "2463",
+  "2507",
+  "2514",
+  "2696",
+  "2728",
+  "2754",
+  "2821",
+  "2884",
+  "2914",
+  "2973",
+  "3018",
+  "3064",
+  "3161",
+  "3172",
+  "3208",
+  "3225",
+  "3248",
+  "3315",
+  "3334",
+  "3358",
+  "3425",
+  "3499",
+  "3537",
+  "3791"
 ]
-ID	1	4079	10	JSON	[
-  "371",
-  "742",
-  "1113",
-  "1484",
-  "1855",
-  "2225",
-  "2596",
-  "2967",
-  "3338",
-  "3709"
+ID	1	4079	50	JSON	[
+  "80",
+  "160",
+  "240",
+  "320",
+  "400",
+  "480",
+  "560",
+  "640",
+  "720",
+  "800",
+  "880",
+  "960",
+  "1040",
+  "1120",
+  "1200",
+  "1280",
+  "1360",
+  "1440",
+  "1520",
+  "1600",
+  "1680",
+  "1760",
+  "1840",
+  "1920",
+  "2000",
+  "2080",
+  "2160",
+  "2240",
+  "2320",
+  "2400",
+  "2480",
+  "2560",
+  "2640",
+  "2720",
+  "2800",
+  "2880",
+  "2960",
+  "3040",
+  "3120",
+  "3200",
+  "3280",
+  "3360",
+  "3440",
+  "3520",
+  "3600",
+  "3680",
+  "3760",
+  "3840",
+  "3920",
+  "4000"
 ]
-Name	A Coruña (La Coruña)	Århus	10	JSON	[
-  "Berdytšiv",
-  "Clermont-Ferrand",
-  "Gatineau",
-  "Itapecerica da Serra",
-  "Kuytun",
-  "Milano",
-  "Ota",
-  "Roanoke",
-  "Split",
-  "Ulsan"
+Name	A Coruña (La Coruña)	Århus	50	JSON	[
+  "Allentown",
+  "Araguari",
+  "Bahtim",
+  "Batangas",
+  "Bialystok",
+  "Brampton",
+  "Calama",
+  "Changchun",
+  "Ciomas",
+  "Cuautla",
+  "Detroit",
+  "Effon-Alaiye",
+  "Firozabad",
+  "Gebze",
+  "Guangyuan",
+  "Hangzhou",
+  "Hradec Králové",
+  "Inazawa",
+  "Jalib al-Shuyukh",
+  "Jubayl",
+  "Kassel",
+  "Kitakyushu",
+  "Kunshan",
+  "Le Mans",
+  "Longueuil",
+  "Malang",
+  "Mati",
+  "Mishan",
+  "Muroran",
+  "Natal",
+  "North York",
+  "Omsk",
+  "Palu",
+  "Phoenix",
+  "Poznan",
+  "Quezon",
+  "Rishra",
+  "Salem",
+  "Sancti-Spíritus",
+  "Sekondi-Takoradi",
+  "Silao",
+  "Stoke-on-Trent",
+  "Taegu",
+  "Teheran",
+  "Tomsk",
+  "Târgu Mures",
+  "Varginha",
+  "Weifang",
+  "Yangjiang",
+  "Zhaodong"
 ]
-Country	ABW	ZWE	10	JSON	[
+Country	ABW	ZWE	50	JSON	[
+  "ARM",
+  "BHS",
   "BRA",
+  "BRB",
+  "BRN",
+  "CAN",
   "CHN",
+  "CIV",
+  "CMR",
+  "COD",
+  "COG",
+  "COL",
+  "DEU",
   "DZA",
+  "ESP",
+  "FRA",
+  "GBR",
   "IDN",
   "IND",
+  "IRL",
+  "IRN",
+  "IRQ",
+  "ISL",
+  "ISR",
+  "ITA",
   "JPN",
-  "MEX",
-  "PHL",
-  "RUS",
-  "USA"
-]
-Population	42	10500000	10	JSON	[
-  "96002",
-  "105819",
-  "118326",
-  "132318",
-  "152397",
-  "181900",
-  "221400",
-  "288173",
-  "398300",
-  "670208"
-]
-Country	ABW	ZWE	10	JSON	[
-  "BGD",
-  "CMR",
-  "ESP",
-  "GUF",
   "KAZ",
+  "KEN",
+  "KOR",
+  "LKA",
+  "MEX",
+  "MHL",
+  "MMR",
+  "NGA",
+  "NZL",
+  "PER",
+  "PHL",
+  "POL",
+  "QAT",
+  "RUS",
+  "RWA",
+  "SAU",
+  "TCD",
+  "TUR",
+  "UKR",
+  "USA",
+  "UZB",
+  "VAT",
+  "VCT",
+  "VNM"
+]
+Population	42	10500000	50	JSON	[
+  "50699",
+  "90601",
+  "92583",
+  "94600",
+  "96800",
+  "98781",
+  "100887",
+  "102681",
+  "105119",
+  "107329",
+  "109500",
+  "112375",
+  "115532",
+  "118718",
+  "121600",
+  "124207",
+  "126820",
+  "130000",
+  "133642",
+  "137776",
+  "142044",
+  "146439",
+  "151060",
+  "157300",
+  "163849",
+  "170428",
+  "176576",
+  "183100",
+  "189569",
+  "197254",
+  "206158",
+  "216735",
+  "227657",
+  "241649",
+  "254842",
+  "270251",
+  "286900",
+  "304477",
+  "326399",
+  "349246",
+  "372840",
+  "409100",
+  "443727",
+  "483155",
+  "540828",
+  "636765",
+  "764902",
+  "993400",
+  "1300977",
+  "2154376"
+]
+Country	ABW	ZWE	50	JSON	[
+  "ALB",
+  "ASM",
+  "AZE",
+  "BFA",
+  "BHS",
+  "BRA",
+  "CAF",
+  "CHE",
+  "CIV",
+  "COD",
+  "CPV",
+  "CZE",
+  "DOM",
+  "EST",
+  "FRA",
+  "GBR",
+  "GIN",
+  "GRC",
+  "HKG",
+  "IDN",
+  "IND",
+  "ISL",
+  "JPN",
+  "KGZ",
+  "LAO",
+  "LIE",
+  "LVA",
   "MDA",
-  "NAM",
-  "PNG",
-  "SVK",
-  "UGA"
+  "MLI",
+  "MNG",
+  "MRT",
+  "MYS",
+  "NER",
+  "NIU",
+  "NRU",
+  "PAN",
+  "PLW",
+  "PYF",
+  "RUS",
+  "SDN",
+  "SLE",
+  "SVN",
+  "TCD",
+  "THA",
+  "TUR",
+  "TZA",
+  "UKR",
+  "UZB",
+  "VNM",
+  "ZAF"
 ]
-Language	Abhyasi	[South]Mande	10	JSON	[
-  "Bariba",
-  "Creole English",
+Language	Abhyasi	[South]Mande	50	JSON	[
+  "Amhara",
+  "Arabic",
+  "Araucan",
+  "Bakhtyari",
+  "Belorussian",
+  "Bubi",
+  "Cebuano",
+  "Chinese",
+  "Comorian",
+  "Creole French",
+  "Danish",
+  "Embera",
   "English",
-  "Georgiana",
+  "Eskimo Languages",
+  "Estonian",
+  "French",
+  "Fries",
+  "Futuna",
+  "German",
+  "Greek",
+  "Hakka",
+  "Hui",
   "Italian",
-  "Lomwe",
-  "Moravian",
-  "Punjabi",
-  "Soga",
-  "Tigrinja"
+  "Joruba",
+  "Kazakh",
+  "Kongo",
+  "Kurdish",
+  "Luchazi",
+  "Makua",
+  "Malinke",
+  "Marshallese",
+  "Mixed Languages",
+  "Nauru",
+  "Nung",
+  "Pangasinan",
+  "Polish",
+  "Portuguese",
+  "Romanian",
+  "Russian",
+  "Sara",
+  "Shona",
+  "Songhai",
+  "Spanish",
+  "Sranantonga",
+  "Tamashek",
+  "Thai",
+  "Tswana",
+  "Tuvalu",
+  "Urdu",
+  "Wolea"
 ]
-Percentage	0.0	99.9	10	JSON	[
+Percentage	0.0	99.9	50	JSON	[
+  "0.0",
+  "0.1",
   "0.2",
+  "0.3",
+  "0.4",
+  "0.5",
   "0.6",
+  "0.7",
+  "0.8",
+  "0.9",
+  "1.0",
+  "1.1",
+  "1.3",
   "1.4",
-  "2.4",
-  "3.9",
-  "6.5",
-  "9.6",
-  "17.7",
-  "45.5",
-  "87.5"
+  "1.6",
+  "1.8",
+  "2.1",
+  "2.3",
+  "2.5",
+  "2.9",
+  "3.2",
+  "3.5",
+  "3.8",
+  "4.4",
+  "4.9",
+  "5.5",
+  "5.9",
+  "6.6",
+  "7.4",
+  "7.8",
+  "8.6",
+  "9.2",
+  "10.3",
+  "11.5",
+  "12.9",
+  "14.6",
+  "17.1",
+  "20.3",
+  "23.8",
+  "31.8",
+  "39.4",
+  "47.5",
+  "55.1",
+  "66.5",
+  "77.2",
+  "86.0",
+  "89.9",
+  "94.3",
+  "97.3",
+  "99.0"
 ]
+explain extended select * from Country where 'Code' between 'BBC' and 'GGG';
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	filtered	Extra
+1	SIMPLE	Country	ALL	NULL	NULL	NULL	NULL	239	100.00	
+Warnings:
+Note	1003	select `world`.`Country`.`Code` AS `Code`,`world`.`Country`.`Name` AS `Name`,`world`.`Country`.`SurfaceArea` AS `SurfaceArea`,`world`.`Country`.`Population` AS `Population`,`world`.`Country`.`Capital` AS `Capital` from `world`.`Country` where 1
+analyze select * from Country where 'Code' between 'BBC' and 'GGG';
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	r_rows	filtered	r_filtered	Extra
+1	SIMPLE	Country	ALL	NULL	NULL	NULL	NULL	239	239.00	100.00	100.00	
 set histogram_type=@save_histogram_type;
 set histogram_size=@save_histogram_size;
 DROP SCHEMA world;
diff --git a/mysql-test/main/statistics_json.test b/mysql-test/main/statistics_json.test
index d5ece058440..4a0f185c28b 100644
--- a/mysql-test/main/statistics_json.test
+++ b/mysql-test/main/statistics_json.test
@@ -12,34 +12,80 @@ set @save_histogram_type=@@histogram_type;
 set @save_histogram_size=@@histogram_size;
 
 CREATE TABLE t1 (
-  a int,
-  b varchar(32),
-  c char(2),
-  d double
+    a int NOT NULL PRIMARY KEY,
+    b varchar(32),
+    c char(16),
+    d date,
+    e double,
+    f bit(3),
+    INDEX idx1 (b, e),
+    INDEX idx2 (c, d),
+    INDEX idx3 (d),
+    INDEX idx4 (e, b, d)
 );
 
 --disable_result_log
-INSERT INTO t1 SELECT seq, seq, seq, seq from seq_1_to_25;
+
+INSERT INTO t1 VALUES
+  (0, NULL, NULL, NULL, NULL, NULL),
+  (7, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'dddddddd', '1990-05-15', 0.1, b'100'),
+  (17, 'vvvvvvvvvvvvv', 'aaaa', '1989-03-12', 0.01, b'101'),
+  (1, 'vvvvvvvvvvvvv', NULL, '1989-03-12', 0.01, b'100'),
+  (12, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1999-07-23', 0.112, b'001'),
+  (23, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.1, b'100'),
+  (8, 'vvvvvvvvvvvvv', 'aaaa', '1999-07-23', 0.1, b'100'),
+  (22, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'aaaa', '1989-03-12', 0.112, b'001'),
+  (31, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'aaaa', '1999-07-23', 0.01, b'001'),
+  (10, NULL, 'aaaa', NULL, 0.01, b'010'),
+  (5, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1999-07-23', 0.1, b'100'),
+  (15, 'vvvvvvvvvvvvv', 'ccccccccc', '1990-05-15', 0.1, b'010'),
+  (30, NULL, 'bbbbbb', NULL, NULL, b'100'),
+  (38, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', NULL, NULL, NULL),
+  (18, 'zzzzzzzzzzzzzzzzzz', 'ccccccccc', '1990-05-15', 0.01, b'010'),
+  (9, 'yyy', 'bbbbbb', '1998-08-28', 0.01, NULL),
+  (29, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.012, b'010'),
+  (3, 'yyy', 'dddddddd',  '1990-05-15', 0.112, b'010'),
+  (39, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', NULL, 0.01, b'100'),
+  (14, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'ccccccccc', '1990-05-15', 0.1, b'100'),
+  (40, 'zzzzzzzzzzzzzzzzzz', 'bbbbbb', '1989-03-12', NULL, NULL),
+  (44, NULL, 'aaaa', '1989-03-12', NULL, b'010'),
+  (19, 'vvvvvvvvvvvvv', 'ccccccccc', '1990-05-15', 0.012, b'011'),
+  (21, 'zzzzzzzzzzzzzzzzzz', 'dddddddd', '1989-03-12', 0.112, b'100'),
+  (45, NULL, NULL, '1989-03-12', NULL, b'011'),
+  (2, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'ccccccccc', '1990-05-15', 0.1, b'001'),
+  (35, 'yyy', 'aaaa', '1990-05-15', 0.05, b'011'),
+  (4, 'vvvvvvvvvvvvv', 'dddddddd', '1999-07-23', 0.01, b'101'),
+  (47, NULL, 'aaaa', '1990-05-15', 0.05, b'010'),
+  (42, NULL, 'ccccccccc', '1989-03-12', 0.01, b'010'),
+  (32, NULL, 'bbbbbb', '1990-05-15', 0.01, b'011'),
+  (49, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww' , 'aaaa', '1990-05-15', NULL, NULL),
+  (43, 'wwwwwwwwwwwwwwwwwwwwwwwwwwww' , 'bbbbbb', '1990-05-15', NULL, b'100'),
+  (37, 'yyy', NULL, '1989-03-12', 0.05, b'011'),
+  (41, 'xxxxxxxxxxxxxxxxxxxxxxxxxx', 'ccccccccc', '1990-05-15', 0.05, NULL),
+  (34, 'yyy', NULL, NULL, NULL, NULL),
+  (33, 'zzzzzzzzzzzzzzzzzz', 'dddddddd', '1989-03-12', 0.05, b'011'),
+  (24,  'wwwwwwwwwwwwwwwwwwwwwwwwwwww', 'dddddddd', '1990-05-15', 0.01,  b'101'),
+  (11, 'yyy', 'ccccccccc', '1999-07-23', 0.1, NULL),
+  (25, 'zzzzzzzzzzzzzzzzzz', 'bbb', '1989-03-12', 0.01,  b'101');
 --enable_result_log
 
 SET histogram_type='JSON';
-# set histogram size to be < row count (25 in this case) to see how histogram behaves
-set histogram_size=10;
+set histogram_size=25;
 
-ANALYZE TABLE t1 PERSISTENT FOR ALL;
-SELECT * FROM mysql.column_stats WHERE table_name='t1';
-explain extended select * from t1 where b between '20' and '70';
+ANALYZE TABLE t1 persistent for all;
+SELECT * FROM mysql.table_stats;
+SELECT * FROM mysql.column_stats;
+SELECT * FROM mysql.index_stats;
+SELECT COUNT(*) FROM t1;
 
-# We then test different valid JSON strings that are invalid histograms.
-UPDATE mysql.column_stats SET histogram='["1", {"a": "b"}, "2"]' WHERE table_name='t1';
-FLUSH TABLES;
---error ER_JSON_HISTOGRAM_PARSE_FAILED
-SELECT * FROM t1;
+explain extended select * from t1 where a between '20' and '70';
+analyze select * from t1 where a between '20' and '70';
 
-UPDATE mysql.column_stats SET histogram='{}' WHERE table_name='t1';
-FLUSH TABLES;
---error ER_JSON_HISTOGRAM_PARSE_FAILED
-SELECT * FROM t1;
+# todo: test different valid JSON strings that are invalid histograms.
+# UPDATE mysql.column_stats SET histogram='["1", {"a": "b"}, "2"]' WHERE table_name='t1';
+# FLUSH TABLES;
+# --error ER_JSON_HISTOGRAM_PARSE_FAILED
+# explain extended select * from t1 where a between '20' and '70';
 
 DELETE FROM mysql.column_stats;
 DROP TABLE t1;
@@ -56,12 +102,14 @@ use world;
 --enable_query_log
 
 set histogram_type='JSON';
-set histogram_size=25;
+set histogram_size=50;
 --disable_result_log
-ANALYZE TABLE Country PERSISTENT FOR ALL;
+ANALYZE TABLE Country, City, CountryLanguage persistent for all;
 --enable_result_log
 
 SELECT column_name, min_value, max_value, hist_size, hist_type, histogram FROM mysql.column_stats;
+explain extended select * from Country where 'Code' between 'BBC' and 'GGG';
+analyze select * from Country where 'Code' between 'BBC' and 'GGG';
 
 set histogram_type=@save_histogram_type;
 set histogram_size=@save_histogram_size;
diff --git a/sql/sql_statistics.cc b/sql/sql_statistics.cc
index b03d537cb2b..5ab90a7587c 100644
--- a/sql/sql_statistics.cc
+++ b/sql/sql_statistics.cc
@@ -1564,9 +1564,9 @@ int Histogram_json::find_bucket(Field *field, const uchar *endpoint)
     mid_val = histogram_bounds[mid];
 
     int res = field->key_cmp((uchar*) mid_val.data(), endpoint);
-    min_bucket_index = mid;
     if (res < 0) {
       low = mid + 1;
+      min_bucket_index = mid;
     } else if (res > 0) {
       high = mid - 1;
     } else {
@@ -1574,6 +1574,9 @@ int Histogram_json::find_bucket(Field *field, const uchar *endpoint)
       break;
     }
   }
+
+  if (min_bucket_index == -1)
+    min_bucket_index = high;
   return min_bucket_index;
 }