YQL-17371: fix ASCII-compatibe bytea output in pgrun (#1356)

10 files changed, 123 insertions, 49 deletions

diff --git a/ydb/library/yql/tests/postgresql/cases/float4.out b/ydb/library/yql/tests/postgresql/cases/float4.out
index 69e63f9ca83..c4128d4416d 100644
--- a/ydb/library/yql/tests/postgresql/cases/float4.out
+++ b/ydb/library/yql/tests/postgresql/cases/float4.out
@@ -320,6 +320,15 @@ SELECT '-9223372036854775808.5'::float4::int8;
 
 SELECT '-9223380000000000000'::float4::int8;
 ERROR:  bigint out of range
+-- Test for correct input rounding in edge cases.
+-- These lists are from Paxson 1991, excluding subnormals and
+-- inputs of over 9 sig. digits.
+SELECT float4send('5e-20'::float4);
+ float4send 
+------------
+ \x1f6c1e4a
+(1 row)
+
 SELECT float4send('67e14'::float4);
  float4send 
 ------------
@@ -350,6 +359,12 @@ SELECT float4send('702990899e-20'::float4);
  \x2cf757ca
 (1 row)
 
+SELECT float4send('3e-23'::float4);
+ float4send 
+------------
+ \x1a111234
+(1 row)
+
 SELECT float4send('57e18'::float4);
  float4send 
 ------------
@@ -392,6 +407,12 @@ SELECT float4send('82381273e-35'::float4);
  \x128289d1
 (1 row)
 
+SELECT float4send('750486563e-38'::float4);
+ float4send 
+------------
+ \x0f18377e
+(1 row)
+
 -- Test that the smallest possible normalized input value inputs
 -- correctly, either in 9-significant-digit or shortest-decimal
 -- format.
diff --git a/ydb/library/yql/tests/postgresql/cases/float4.sql b/ydb/library/yql/tests/postgresql/cases/float4.sql
index 7260ea8a372..8a6adaf11d6 100644
--- a/ydb/library/yql/tests/postgresql/cases/float4.sql
+++ b/ydb/library/yql/tests/postgresql/cases/float4.sql
@@ -76,11 +76,16 @@ SELECT '9223369837831520256'::float4::int8;
 SELECT '9223372036854775807'::float4::int8;
 SELECT '-9223372036854775808.5'::float4::int8;
 SELECT '-9223380000000000000'::float4::int8;
+-- Test for correct input rounding in edge cases.
+-- These lists are from Paxson 1991, excluding subnormals and
+-- inputs of over 9 sig. digits.
+SELECT float4send('5e-20'::float4);
 SELECT float4send('67e14'::float4);
 SELECT float4send('985e15'::float4);
 SELECT float4send('55895e-16'::float4);
 SELECT float4send('7038531e-32'::float4);
 SELECT float4send('702990899e-20'::float4);
+SELECT float4send('3e-23'::float4);
 SELECT float4send('57e18'::float4);
 SELECT float4send('789e-35'::float4);
 SELECT float4send('2539e-18'::float4);
@@ -88,6 +93,7 @@ SELECT float4send('76173e28'::float4);
 SELECT float4send('887745e-11'::float4);
 SELECT float4send('5382571e-37'::float4);
 SELECT float4send('82381273e-35'::float4);
+SELECT float4send('750486563e-38'::float4);
 -- Test that the smallest possible normalized input value inputs
 -- correctly, either in 9-significant-digit or shortest-decimal
 -- format.
diff --git a/ydb/library/yql/tests/postgresql/cases/float8.out b/ydb/library/yql/tests/postgresql/cases/float8.out
index 87f089819ce..69b9840413c 100644
--- a/ydb/library/yql/tests/postgresql/cases/float8.out
+++ b/ydb/library/yql/tests/postgresql/cases/float8.out
@@ -24,6 +24,13 @@ SELECT '-10e-400'::float8;
 ERROR:  "-10e-400" is out of range for type double precision
 LINE 1: SELECT '-10e-400'::float8;
                ^
+-- test smallest normalized input
+SELECT float8send('2.2250738585072014E-308'::float8);
+     float8send     
+--------------------
+ \x0010000000000000
+(1 row)
+
 -- bad input
 INSERT INTO FLOAT8_TBL(f1) VALUES ('');
 ERROR:  invalid input syntax for type double precision: ""
diff --git a/ydb/library/yql/tests/postgresql/cases/float8.sql b/ydb/library/yql/tests/postgresql/cases/float8.sql
index d45889997f9..98b8d6e05a7 100644
--- a/ydb/library/yql/tests/postgresql/cases/float8.sql
+++ b/ydb/library/yql/tests/postgresql/cases/float8.sql
@@ -12,6 +12,8 @@ SELECT '10e400'::float8;
 SELECT '-10e400'::float8;
 SELECT '10e-400'::float8;
 SELECT '-10e-400'::float8;
+-- test smallest normalized input
+SELECT float8send('2.2250738585072014E-308'::float8);
 -- bad input
 INSERT INTO FLOAT8_TBL(f1) VALUES ('');
 INSERT INTO FLOAT8_TBL(f1) VALUES ('     ');
diff --git a/ydb/library/yql/tests/postgresql/cases/strings.out b/ydb/library/yql/tests/postgresql/cases/strings.out
index c0a3286c08b..f5c88263fc9 100644
--- a/ydb/library/yql/tests/postgresql/cases/strings.out
+++ b/ydb/library/yql/tests/postgresql/cases/strings.out
@@ -178,6 +178,30 @@ SELECT E'\\xDe00BeEf'::bytea;
  \xde00beef
 (1 row)
 
+SELECT E'DeAdBeEf'::bytea;
+       bytea        
+--------------------
+ \x4465416442654566
+(1 row)
+
+SELECT E'De\\000dBeEf'::bytea;
+       bytea        
+--------------------
+ \x4465006442654566
+(1 row)
+
+SELECT E'De\123dBeEf'::bytea;
+       bytea        
+--------------------
+ \x4465536442654566
+(1 row)
+
+SELECT E'De\\123dBeEf'::bytea;
+       bytea        
+--------------------
+ \x4465536442654566
+(1 row)
+
 SELECT E'De\\678dBeEf'::bytea;
 ERROR:  invalid input syntax for type bytea
 LINE 1: SELECT E'De\\678dBeEf'::bytea;
@@ -207,6 +231,12 @@ SELECT E'DeAdBeEf'::bytea;
  DeAdBeEf
 (1 row)
 
+SELECT E'De\\000dBeEf'::bytea;
+    bytea    
+-------------
+ De\000dBeEf
+(1 row)
+
 SELECT E'De\\123dBeEf'::bytea;
   bytea   
 ----------
diff --git a/ydb/library/yql/tests/postgresql/cases/strings.sql b/ydb/library/yql/tests/postgresql/cases/strings.sql
index 2e90a9bd00e..ce760be418c 100644
--- a/ydb/library/yql/tests/postgresql/cases/strings.sql
+++ b/ydb/library/yql/tests/postgresql/cases/strings.sql
@@ -51,12 +51,17 @@ SELECT E'\\x De Ad Be Ef '::bytea;
 SELECT E'\\xDeAdBeE'::bytea;
 SELECT E'\\xDeAdBeEx'::bytea;
 SELECT E'\\xDe00BeEf'::bytea;
+SELECT E'DeAdBeEf'::bytea;
+SELECT E'De\\000dBeEf'::bytea;
+SELECT E'De\123dBeEf'::bytea;
+SELECT E'De\\123dBeEf'::bytea;
 SELECT E'De\\678dBeEf'::bytea;
 SET bytea_output TO escape;
 SELECT E'\\xDeAdBeEf'::bytea;
 SELECT E'\\x De Ad Be Ef '::bytea;
 SELECT E'\\xDe00BeEf'::bytea;
 SELECT E'DeAdBeEf'::bytea;
+SELECT E'De\\000dBeEf'::bytea;
 SELECT E'De\\123dBeEf'::bytea;
 SELECT CAST(name 'namefield' AS text) AS "text(name)";
 SELECT CAST(name 'namefield' AS char(10)) AS "char(name)";
diff --git a/ydb/library/yql/tests/postgresql/pg_tests.csv b/ydb/library/yql/tests/postgresql/pg_tests.csv
index 42b87b603dc..fe596ddc218 100644
--- a/ydb/library/yql/tests/postgresql/pg_tests.csv
+++ b/ydb/library/yql/tests/postgresql/pg_tests.csv
@@ -13,8 +13,8 @@ date,264,200,75.76
 dbsize,24,24,100.0
 delete,10,0,0.0
 expressions,63,14,22.22
-float4,96,80,83.33
-float8,168,132,78.57
+float4,96,82,85.42
+float8,168,140,83.33
 functional_deps,40,7,17.5
 horology,306,79,25.82
 insert,357,15,4.2
@@ -40,7 +40,7 @@ select_distinct_on,4,0,0.0
 select_having,23,16,69.57
 select_implicit,44,13,29.55
 select_into,67,3,4.48
-strings,390,81,20.77
+strings,390,86,22.05
 subselect,234,5,2.14
 text,76,16,21.05
 time,39,33,84.62
diff --git a/ydb/library/yql/tests/postgresql/status.md b/ydb/library/yql/tests/postgresql/status.md
index dcc3eb0e158..320d7411326 100644
--- a/ydb/library/yql/tests/postgresql/status.md
+++ b/ydb/library/yql/tests/postgresql/status.md
@@ -11,12 +11,12 @@
 || 7 | int4 | 70 | 70 | 100.0 | 29.09.2023 | ||
 || 8 | int8 | 142 | 113 (+65) | 79.58 | 22.01.2024 | YQL-17614 ||
 || 9 | oid | 27 | 27 (+6) | 100.0 | 22.01.2024 | YQL-17623 ||
-|| 10 | float4 | 96 | 80 (+32) | 83.33 | 23.01.2024 | YQL-17586 ||
-|| 11 | float8 | 168 | 132 (+36) | 78.57 | 23.01.2024 | YQL-17628 ||
+|| 10 | float4 | 96 | 82 (+2) | 85.42 | 26.01.2024 | YQL-17586 ||
+|| 11 | float8 | 168 | 140 (+8) | 83.33 | 26.01.2024 | YQL-17628 ||
 || 12 | bit | 115 | 84 | 73.04 | 12.12.2023 | YQL-17634 ||
 || 13 | numeric | 915 | 721 (+6) | 78.8 | 23.01.2024 | YQL-17629 ||
 || 14 | uuid | 36 | 0 | 0.0 | 02.05.2023 | YQL-17636 ||
-|| 15 | strings | 390 | 81 (+50) | 20.77 | 24.01.2024 | YQL-17587 ||
+|| 15 | strings | 390 | 86 (+5) | 22.05 | 26.01.2024 | YQL-17587 ||
 || 16 | numerology | 24 | 8 | 33.33 | 26.07.2023 |  ||
 || 17 | date | 264 | 200 | 75.76 | 12.12.2023 | ||
 || 18 | time | 39 | 33 | 84.62 | 12.12.2023 | ||
diff --git a/ydb/library/yql/tests/postgresql/status.old b/ydb/library/yql/tests/postgresql/status.old
index f9f1d0a21e6..dcc3eb0e158 100644
--- a/ydb/library/yql/tests/postgresql/status.old
+++ b/ydb/library/yql/tests/postgresql/status.old
@@ -11,12 +11,12 @@
 || 7 | int4 | 70 | 70 | 100.0 | 29.09.2023 | ||
 || 8 | int8 | 142 | 113 (+65) | 79.58 | 22.01.2024 | YQL-17614 ||
 || 9 | oid | 27 | 27 (+6) | 100.0 | 22.01.2024 | YQL-17623 ||
-|| 10 | float4 | 96 | 80 (+14) | 83.33 | 23.01.2024 | YQL-17586 ||
+|| 10 | float4 | 96 | 80 (+32) | 83.33 | 23.01.2024 | YQL-17586 ||
 || 11 | float8 | 168 | 132 (+36) | 78.57 | 23.01.2024 | YQL-17628 ||
 || 12 | bit | 115 | 84 | 73.04 | 12.12.2023 | YQL-17634 ||
 || 13 | numeric | 915 | 721 (+6) | 78.8 | 23.01.2024 | YQL-17629 ||
 || 14 | uuid | 36 | 0 | 0.0 | 02.05.2023 | YQL-17636 ||
-|| 15 | strings | 390 | 76 (+45) | 19.49 | 23.01.2024 | YQL-17587 ||
+|| 15 | strings | 390 | 81 (+50) | 20.77 | 24.01.2024 | YQL-17587 ||
 || 16 | numerology | 24 | 8 | 33.33 | 26.07.2023 |  ||
 || 17 | date | 264 | 200 | 75.76 | 12.12.2023 | ||
 || 18 | time | 39 | 33 | 84.62 | 12.12.2023 | ||
diff --git a/ydb/library/yql/tools/pgrun/pgrun.cpp b/ydb/library/yql/tools/pgrun/pgrun.cpp
index ff1b632fc68..29dfd266f1e 100644
--- a/ydb/library/yql/tools/pgrun/pgrun.cpp
+++ b/ydb/library/yql/tools/pgrun/pgrun.cpp
@@ -815,54 +815,57 @@ std::string FormatCell(const TString& data, const TColumn& column, size_t index,
 
 TString GetCellData(const NYT::TNode& cell, const TColumn& column) {
     if (column.Type == "bytea") {
-        if (cell.IsList()) {
-            TString result;
+        const auto rawValue = (cell.IsList())
+            ? Base64Decode(cell.AsList()[0].AsString())
+            : cell.AsString();
 
-            const auto rawValue = Base64Decode(cell.AsList()[0].AsString());
-            switch (byteaOutput) {
-                case EByteaOutput::hex: {
-                    const auto expectedSize = rawValue.size() * 2 + 2;
-                    result.resize(expectedSize);
-                    result[0] = '\\';
-                    result[1] = 'x';
-                    const auto cnt = HexEncode(rawValue.data(), rawValue.size(), result.begin() + 2);
+        switch (byteaOutput) {
+            case EByteaOutput::hex: {
+                TString result;
 
-                    Y_ASSERT(cnt + 2 == expectedSize);
+                const auto expectedSize = rawValue.size() * 2 + 2;
+                result.resize(expectedSize);
+                result[0] = '\\';
+                result[1] = 'x';
+                const auto cnt = HexEncode(rawValue.data(), rawValue.size(), result.begin() + 2);
 
-                    return result;
-                }
-                case EByteaOutput::escape: {
-                    ui64 expectedSize = std::accumulate(rawValue.cbegin(), rawValue.cend(), 0U,
-                        [] (ui64 acc, char c) {
-                            return acc + ((c == '\\')
-                                          ? 2
-                                          : ((ui8)c < 0x20 || 0x7e < (ui8)c)
-                                            ? 4
-                                            : 1);
-                        });
-                    result.resize(expectedSize);
-                    auto p = result.begin();
-                    for (const auto c : rawValue) {
-                        if (c == '\\') {
-                            *p++ = '\\';
-                            *p++ = '\\';
-                        } else if ((ui8)c < 0x20 || 0x7e < (ui8)c) {
-                            auto val = (ui8)c;
+                Y_ASSERT(cnt + 2 == expectedSize);
 
-                            *p++ = '\\';
-                            *p++ = ((val >> 6) & 03) + '0';
-                            *p++ = ((val >> 3) & 07) + '0';
-                            *p++ =  (val & 07) + '0';
-                        } else {
-                            *p++ = c;
-                        }
-                    }
+                return result;
+            }
+            case EByteaOutput::escape: {
+                TString result;
+
+                ui64 expectedSize = std::accumulate(rawValue.cbegin(), rawValue.cend(), 0U,
+                    [] (ui64 acc, char c) {
+                        return acc + ((c == '\\')
+                                        ? 2
+                                        : ((ui8)c < 0x20 || 0x7e < (ui8)c)
+                                        ? 4
+                                        : 1);
+                    });
+                result.resize(expectedSize);
+                auto p = result.begin();
+                for (const auto c : rawValue) {
+                    if (c == '\\') {
+                        *p++ = '\\';
+                        *p++ = '\\';
+                    } else if ((ui8)c < 0x20 || 0x7e < (ui8)c) {
+                        auto val = (ui8)c;
 
-                    return result;
+                        *p++ = '\\';
+                        *p++ = ((val >> 6) & 03) + '0';
+                        *p++ = ((val >> 3) & 07) + '0';
+                        *p++ =  (val & 07) + '0';
+                    } else {
+                        *p++ = c;
+                    }
                 }
-                default:
-                    throw yexception() << "Unhandled EByteaOutput value";
+
+                return result;
             }
+            default:
+                throw yexception() << "Unhandled EByteaOutput value";
         }
     }
     return cell.AsString();