This is an automated email from the ASF dual-hosted git repository.
yangzhg pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-doris.git
The following commit(s) were added to refs/heads/master by this push:
new 16020cb [fix](lateral-view) Fix bug that explode_json_array_string
return unstable result (#8152)
16020cb is described below
commit 16020cbdf9d1ec685514380b63eb59a8897477da
Author: Mingyu Chen <morningman....@gmail.com>
AuthorDate: Mon Feb 21 09:38:36 2022 +0800
[fix](lateral-view) Fix bug that explode_json_array_string return unstable
result (#8152)
Co-authored-by: morningman <chenmin...@baidu.com>
---
be/src/exprs/table_function/explode_json_array.cpp | 17 +++++++++++------
be/src/exprs/table_function/explode_json_array.h | 1 -
thirdparty/download-thirdparty.sh | 8 +++++---
3 files changed, 16 insertions(+), 10 deletions(-)
diff --git a/be/src/exprs/table_function/explode_json_array.cpp
b/be/src/exprs/table_function/explode_json_array.cpp
index e82c9ec..43fb1f3 100644
--- a/be/src/exprs/table_function/explode_json_array.cpp
+++ b/be/src/exprs/table_function/explode_json_array.cpp
@@ -68,8 +68,10 @@ int ParsedData::set_output(ExplodeJsonArrayType type,
rapidjson::Document& docum
switch (v.GetType()) {
case rapidjson::Type::kStringType:
_backup_string.emplace_back(v.GetString(),
v.GetStringLength());
- _data_string.emplace_back(_backup_string.back());
_string_nulls.push_back(false);
+ // do not set _data_string here.
+ // Because the address of the string stored in
`_backup_string` may
+ // change each time `emplace_back()` is called.
break;
case rapidjson::Type::kNumberType:
if (v.IsUint()) {
@@ -84,8 +86,10 @@ int ParsedData::set_output(ExplodeJsonArrayType type,
rapidjson::Document& docum
wbytes = sprintf(tmp_buf, "%f", v.GetDouble());
}
_backup_string.emplace_back(tmp_buf, wbytes);
- _data_string.emplace_back(_backup_string.back());
_string_nulls.push_back(false);
+ // do not set _data_string here.
+ // Because the address of the string stored in
`_backup_string` may
+ // change each time `emplace_back()` is called.
break;
case rapidjson::Type::kFalseType:
_data_string.emplace_back(true_value);
@@ -106,6 +110,11 @@ int ParsedData::set_output(ExplodeJsonArrayType type,
rapidjson::Document& docum
}
++i;
}
+ // Must set _data_string at the end, so that we can
+ // save the real addr of string in `_backup_string` to
`_data_string`.
+ for (auto& str : _backup_string) {
+ _data_string.emplace_back(str);
+ }
break;
}
default:
@@ -139,21 +148,17 @@ Status ExplodeJsonArrayTableFunction::process(TupleRow*
tuple_row) {
StringVal text =
_expr_context->root()->get_child(0)->get_string_val(_expr_context, tuple_row);
if (text.is_null || text.len == 0) {
- // _set_null_output();
_is_current_empty = true;
} else {
rapidjson::Document document;
document.Parse((char*) text.ptr, text.len);
if (UNLIKELY(document.HasParseError()) || !document.IsArray() ||
document.GetArray().Size() == 0) {
- // _set_null_output();
_is_current_empty = true;
} else {
_cur_size = _parsed_data.set_output(_type, document);
_cur_offset = 0;
- // _eos = _cur_size == 0;
}
}
- // _is_current_empty = _eos;
return Status::OK();
}
diff --git a/be/src/exprs/table_function/explode_json_array.h
b/be/src/exprs/table_function/explode_json_array.h
index ae5c657..8ecbd2d 100644
--- a/be/src/exprs/table_function/explode_json_array.h
+++ b/be/src/exprs/table_function/explode_json_array.h
@@ -91,7 +91,6 @@ struct ParsedData {
*output = _data[offset];
break;
case ExplodeJsonArrayType::STRING:
- // LOG(INFO) << "cmy get_value offset: " << offset << ", is
null: " << _string_nulls[offset] << ", data: " << (_string_nulls[offset] ?
"null2" : _backup_string[offset]);
*output = _string_nulls[offset] ? nullptr :
&_data_string[offset];
break;
default:
diff --git a/thirdparty/download-thirdparty.sh
b/thirdparty/download-thirdparty.sh
index c6650bd..768e92d 100755
--- a/thirdparty/download-thirdparty.sh
+++ b/thirdparty/download-thirdparty.sh
@@ -72,6 +72,8 @@ md5sum_func() {
return 0
}
+# return 0 if download succeed.
+# return 1 if not.
download_func() {
local FILENAME=$1
local DOWNLOAD_URL=$2
@@ -134,18 +136,18 @@ do
if test "x$REPOSITORY_URL" = x; then
URL=$TP_ARCH"_DOWNLOAD"
download_func ${!NAME} ${!URL} $TP_SOURCE_DIR ${!MD5SUM}
- if [ "$?"x == "0"x ]; then
+ if [ "$?"x != "0"x ]; then
echo "Failed to download ${!NAME}"
exit 1
fi
else
URL="${REPOSITORY_URL}/${!NAME}"
download_func ${!NAME} ${URL} $TP_SOURCE_DIR ${!MD5SUM}
- if [ "$?x" == "0x" ]; then
+ if [ "$?x" != "0x" ]; then
#try to download from home
URL=$TP_ARCH"_DOWNLOAD"
download_func ${!NAME} ${!URL} $TP_SOURCE_DIR ${!MD5SUM}
- if [ "$?x" == "1x" ]; then
+ if [ "$?"x != "0"x ]; then
echo "Failed to download ${!NAME}"
exit 1 # download failed again exit.
fi
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
