fluent-plugin-aliyun-odps 0.0.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +8 -0
- data/CHANGELOG.md +0 -0
- data/Gemfile +3 -0
- data/License +13 -0
- data/README.md +120 -0
- data/Rakefile +14 -0
- data/VERSION +1 -0
- data/fluent-plugin-aliyun-odps.gemspec +27 -0
- data/lib/fluent/plugin/conf/config.rb +32 -0
- data/lib/fluent/plugin/digest/crc.rb +120 -0
- data/lib/fluent/plugin/digest/crc32.rb +125 -0
- data/lib/fluent/plugin/digest/crc32c.rb +105 -0
- data/lib/fluent/plugin/exceptions.rb +49 -0
- data/lib/fluent/plugin/http/http_connection.rb +130 -0
- data/lib/fluent/plugin/http/http_flag.rb +73 -0
- data/lib/fluent/plugin/odps/odps_table.rb +253 -0
- data/lib/fluent/plugin/odps/odps_table_schema.rb +64 -0
- data/lib/fluent/plugin/odps/xml_template.rb +57 -0
- data/lib/fluent/plugin/odps/xstream_pack.pb.rb +21 -0
- data/lib/fluent/plugin/odps/xstream_pack.proto +8 -0
- data/lib/fluent/plugin/out_odps.rb +373 -0
- data/lib/fluent/plugin/serialize/deserializer.rb +32 -0
- data/lib/fluent/plugin/serialize/serializer.rb +141 -0
- data/lib/fluent/plugin/stream_client.rb +111 -0
- data/lib/fluent/plugin/stream_reader.rb +53 -0
- data/lib/fluent/plugin/stream_writer.rb +152 -0
- data/odps_example.conf +30 -0
- metadata +174 -0
@@ -0,0 +1,105 @@
|
|
1
|
+
#
|
2
|
+
#Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
#or more contributor license agreements. See the NOTICE file
|
4
|
+
#distributed with this work for additional information
|
5
|
+
#regarding copyright ownership. The ASF licenses this file
|
6
|
+
#to you under the Apache License, Version 2.0 (the
|
7
|
+
#"License"); you may not use this file except in compliance
|
8
|
+
#with the License. You may obtain a copy of the License at
|
9
|
+
#
|
10
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
#
|
12
|
+
#Unless required by applicable law or agreed to in writing,
|
13
|
+
#software distributed under the License is distributed on an
|
14
|
+
#"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
#KIND, either express or implied. See the License for the
|
16
|
+
#specific language governing permissions and limitations
|
17
|
+
#under the License.
|
18
|
+
#
|
19
|
+
require_relative 'crc32'
|
20
|
+
module Digest
|
21
|
+
#
|
22
|
+
# Implements the CRC32c algorithm.
|
23
|
+
#
|
24
|
+
class CRC32c < CRC32
|
25
|
+
# Generated by `./pycrc.py --algorithm=table-driven --model=crc-32c --generate=c`
|
26
|
+
TABLE = [
|
27
|
+
0x00000000, 0xf26b8303, 0xe13b70f7, 0x1350f3f4,
|
28
|
+
0xc79a971f, 0x35f1141c, 0x26a1e7e8, 0xd4ca64eb,
|
29
|
+
0x8ad958cf, 0x78b2dbcc, 0x6be22838, 0x9989ab3b,
|
30
|
+
0x4d43cfd0, 0xbf284cd3, 0xac78bf27, 0x5e133c24,
|
31
|
+
0x105ec76f, 0xe235446c, 0xf165b798, 0x030e349b,
|
32
|
+
0xd7c45070, 0x25afd373, 0x36ff2087, 0xc494a384,
|
33
|
+
0x9a879fa0, 0x68ec1ca3, 0x7bbcef57, 0x89d76c54,
|
34
|
+
0x5d1d08bf, 0xaf768bbc, 0xbc267848, 0x4e4dfb4b,
|
35
|
+
0x20bd8ede, 0xd2d60ddd, 0xc186fe29, 0x33ed7d2a,
|
36
|
+
0xe72719c1, 0x154c9ac2, 0x061c6936, 0xf477ea35,
|
37
|
+
0xaa64d611, 0x580f5512, 0x4b5fa6e6, 0xb93425e5,
|
38
|
+
0x6dfe410e, 0x9f95c20d, 0x8cc531f9, 0x7eaeb2fa,
|
39
|
+
0x30e349b1, 0xc288cab2, 0xd1d83946, 0x23b3ba45,
|
40
|
+
0xf779deae, 0x05125dad, 0x1642ae59, 0xe4292d5a,
|
41
|
+
0xba3a117e, 0x4851927d, 0x5b016189, 0xa96ae28a,
|
42
|
+
0x7da08661, 0x8fcb0562, 0x9c9bf696, 0x6ef07595,
|
43
|
+
0x417b1dbc, 0xb3109ebf, 0xa0406d4b, 0x522bee48,
|
44
|
+
0x86e18aa3, 0x748a09a0, 0x67dafa54, 0x95b17957,
|
45
|
+
0xcba24573, 0x39c9c670, 0x2a993584, 0xd8f2b687,
|
46
|
+
0x0c38d26c, 0xfe53516f, 0xed03a29b, 0x1f682198,
|
47
|
+
0x5125dad3, 0xa34e59d0, 0xb01eaa24, 0x42752927,
|
48
|
+
0x96bf4dcc, 0x64d4cecf, 0x77843d3b, 0x85efbe38,
|
49
|
+
0xdbfc821c, 0x2997011f, 0x3ac7f2eb, 0xc8ac71e8,
|
50
|
+
0x1c661503, 0xee0d9600, 0xfd5d65f4, 0x0f36e6f7,
|
51
|
+
0x61c69362, 0x93ad1061, 0x80fde395, 0x72966096,
|
52
|
+
0xa65c047d, 0x5437877e, 0x4767748a, 0xb50cf789,
|
53
|
+
0xeb1fcbad, 0x197448ae, 0x0a24bb5a, 0xf84f3859,
|
54
|
+
0x2c855cb2, 0xdeeedfb1, 0xcdbe2c45, 0x3fd5af46,
|
55
|
+
0x7198540d, 0x83f3d70e, 0x90a324fa, 0x62c8a7f9,
|
56
|
+
0xb602c312, 0x44694011, 0x5739b3e5, 0xa55230e6,
|
57
|
+
0xfb410cc2, 0x092a8fc1, 0x1a7a7c35, 0xe811ff36,
|
58
|
+
0x3cdb9bdd, 0xceb018de, 0xdde0eb2a, 0x2f8b6829,
|
59
|
+
0x82f63b78, 0x709db87b, 0x63cd4b8f, 0x91a6c88c,
|
60
|
+
0x456cac67, 0xb7072f64, 0xa457dc90, 0x563c5f93,
|
61
|
+
0x082f63b7, 0xfa44e0b4, 0xe9141340, 0x1b7f9043,
|
62
|
+
0xcfb5f4a8, 0x3dde77ab, 0x2e8e845f, 0xdce5075c,
|
63
|
+
0x92a8fc17, 0x60c37f14, 0x73938ce0, 0x81f80fe3,
|
64
|
+
0x55326b08, 0xa759e80b, 0xb4091bff, 0x466298fc,
|
65
|
+
0x1871a4d8, 0xea1a27db, 0xf94ad42f, 0x0b21572c,
|
66
|
+
0xdfeb33c7, 0x2d80b0c4, 0x3ed04330, 0xccbbc033,
|
67
|
+
0xa24bb5a6, 0x502036a5, 0x4370c551, 0xb11b4652,
|
68
|
+
0x65d122b9, 0x97baa1ba, 0x84ea524e, 0x7681d14d,
|
69
|
+
0x2892ed69, 0xdaf96e6a, 0xc9a99d9e, 0x3bc21e9d,
|
70
|
+
0xef087a76, 0x1d63f975, 0x0e330a81, 0xfc588982,
|
71
|
+
0xb21572c9, 0x407ef1ca, 0x532e023e, 0xa145813d,
|
72
|
+
0x758fe5d6, 0x87e466d5, 0x94b49521, 0x66df1622,
|
73
|
+
0x38cc2a06, 0xcaa7a905, 0xd9f75af1, 0x2b9cd9f2,
|
74
|
+
0xff56bd19, 0x0d3d3e1a, 0x1e6dcdee, 0xec064eed,
|
75
|
+
0xc38d26c4, 0x31e6a5c7, 0x22b65633, 0xd0ddd530,
|
76
|
+
0x0417b1db, 0xf67c32d8, 0xe52cc12c, 0x1747422f,
|
77
|
+
0x49547e0b, 0xbb3ffd08, 0xa86f0efc, 0x5a048dff,
|
78
|
+
0x8ecee914, 0x7ca56a17, 0x6ff599e3, 0x9d9e1ae0,
|
79
|
+
0xd3d3e1ab, 0x21b862a8, 0x32e8915c, 0xc083125f,
|
80
|
+
0x144976b4, 0xe622f5b7, 0xf5720643, 0x07198540,
|
81
|
+
0x590ab964, 0xab613a67, 0xb831c993, 0x4a5a4a90,
|
82
|
+
0x9e902e7b, 0x6cfbad78, 0x7fab5e8c, 0x8dc0dd8f,
|
83
|
+
0xe330a81a, 0x115b2b19, 0x020bd8ed, 0xf0605bee,
|
84
|
+
0x24aa3f05, 0xd6c1bc06, 0xc5914ff2, 0x37faccf1,
|
85
|
+
0x69e9f0d5, 0x9b8273d6, 0x88d28022, 0x7ab90321,
|
86
|
+
0xae7367ca, 0x5c18e4c9, 0x4f48173d, 0xbd23943e,
|
87
|
+
0xf36e6f75, 0x0105ec76, 0x12551f82, 0xe03e9c81,
|
88
|
+
0x34f4f86a, 0xc69f7b69, 0xd5cf889d, 0x27a40b9e,
|
89
|
+
0x79b737ba, 0x8bdcb4b9, 0x988c474d, 0x6ae7c44e,
|
90
|
+
0xbe2da0a5, 0x4c4623a6, 0x5f16d052, 0xad7d5351
|
91
|
+
]
|
92
|
+
#
|
93
|
+
# Updates the CRC32 checksum.
|
94
|
+
#
|
95
|
+
# @param [String] data
|
96
|
+
# The data to update the checksum with.
|
97
|
+
#
|
98
|
+
def update(data)
|
99
|
+
data.each_byte do |b|
|
100
|
+
@crc = (((@crc >> 8) & 0x00ffffff) ^ TABLE[(@crc ^ b) & 0xff])
|
101
|
+
end
|
102
|
+
return self
|
103
|
+
end
|
104
|
+
end
|
105
|
+
end
|
@@ -0,0 +1,49 @@
|
|
1
|
+
#
|
2
|
+
#Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
#or more contributor license agreements. See the NOTICE file
|
4
|
+
#distributed with this work for additional information
|
5
|
+
#regarding copyright ownership. The ASF licenses this file
|
6
|
+
#to you under the Apache License, Version 2.0 (the
|
7
|
+
#"License"); you may not use this file except in compliance
|
8
|
+
#with the License. You may obtain a copy of the License at
|
9
|
+
#
|
10
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
#
|
12
|
+
#Unless required by applicable law or agreed to in writing,
|
13
|
+
#software distributed under the License is distributed on an
|
14
|
+
#"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
#KIND, either express or implied. See the License for the
|
16
|
+
#specific language governing permissions and limitations
|
17
|
+
#under the License.
|
18
|
+
#
|
19
|
+
module OdpsDatahub
|
20
|
+
$INVALID_ARGUMENT = "InvalidArgument"
|
21
|
+
$SCHEMA_NOT_MATCH = "SchemaNotMatch"
|
22
|
+
$PACK_SIZE_EXCEED = "PackSizeExceed"
|
23
|
+
$ACCESS_DENIED = "AccessDenied"
|
24
|
+
|
25
|
+
class OdpsDatahubException < StandardError
|
26
|
+
attr_reader :mMessage, :mErrorCode, :mRequestId
|
27
|
+
def initialize(code, msg, reqId = "")
|
28
|
+
@mMessage = msg
|
29
|
+
@mErrorCode = code
|
30
|
+
@mRequestId = reqId
|
31
|
+
end
|
32
|
+
|
33
|
+
def getMessage
|
34
|
+
@mMessage
|
35
|
+
end
|
36
|
+
|
37
|
+
def getCode
|
38
|
+
@mErrorCode
|
39
|
+
end
|
40
|
+
|
41
|
+
def getRequestId
|
42
|
+
@mRequestId
|
43
|
+
end
|
44
|
+
|
45
|
+
def to_s
|
46
|
+
return "ErrorCode: " + @mErrorCode + ", Message: " + @mMessage + ", requestId:" + @mRequestId
|
47
|
+
end
|
48
|
+
end
|
49
|
+
end
|
@@ -0,0 +1,130 @@
|
|
1
|
+
#
|
2
|
+
#Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
#or more contributor license agreements. See the NOTICE file
|
4
|
+
#distributed with this work for additional information
|
5
|
+
#regarding copyright ownership. The ASF licenses this file
|
6
|
+
#to you under the Apache License, Version 2.0 (the
|
7
|
+
#"License"); you may not use this file except in compliance
|
8
|
+
#with the License. You may obtain a copy of the License at
|
9
|
+
#
|
10
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
#
|
12
|
+
#Unless required by applicable law or agreed to in writing,
|
13
|
+
#software distributed under the License is distributed on an
|
14
|
+
#"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
#KIND, either express or implied. See the License for the
|
16
|
+
#specific language governing permissions and limitations
|
17
|
+
#under the License.
|
18
|
+
#
|
19
|
+
require 'net/http'
|
20
|
+
require 'base64'
|
21
|
+
require 'openssl'
|
22
|
+
require_relative '../exceptions'
|
23
|
+
require_relative '../conf/config'
|
24
|
+
require_relative 'http_flag'
|
25
|
+
Net::HTTP.version_1_2
|
26
|
+
|
27
|
+
module OdpsDatahub
|
28
|
+
class HttpConnection
|
29
|
+
attr_reader :mHeader, :mParam, :mUri, :mReq, :mMethod, :mPath, :mStream, :mOdpsConfig
|
30
|
+
def initialize(odpsConfig, headers, params, path, method, stream = nil, isodpsurl = false)
|
31
|
+
@mOdpsConfig = odpsConfig
|
32
|
+
@mHeader = headers
|
33
|
+
@mParam = params
|
34
|
+
@mPath = path
|
35
|
+
@mMethod = method
|
36
|
+
@mStream = stream
|
37
|
+
@mIsOdpsUrl = isodpsurl
|
38
|
+
buildRequest
|
39
|
+
end
|
40
|
+
|
41
|
+
def buildRequest
|
42
|
+
path = ""
|
43
|
+
separater = '?'
|
44
|
+
@mParam.each { |key , value|
|
45
|
+
if value != ""
|
46
|
+
path += separater + key.to_s + '=' + value.to_s
|
47
|
+
else
|
48
|
+
path += separater + key.to_s
|
49
|
+
end
|
50
|
+
separater = '&'
|
51
|
+
}
|
52
|
+
if @mIsOdpsUrl
|
53
|
+
@mUri = URI.parse(@mOdpsConfig.odpsEndpoint + @mPath + path)
|
54
|
+
else
|
55
|
+
@mUri = URI.parse(@mOdpsConfig.datahubEndpoint + @mPath + path)
|
56
|
+
end
|
57
|
+
|
58
|
+
if !@mHeader.has_key?($CONTENT_MD5)
|
59
|
+
@mHeader[$CONTENT_MD5] = ""
|
60
|
+
end
|
61
|
+
if !@mHeader.has_key?($CONTENT_TYPE)
|
62
|
+
@mHeader[$CONTENT_TYPE] = ""
|
63
|
+
end
|
64
|
+
@mHeader[$DATE] = Time.now.utc.strftime("%a, %d %b %Y %H:%M:%S GMT")
|
65
|
+
if not @mIsOdpsUrl
|
66
|
+
@mHeader[$TUNNEL_STREAM_VERSION] = "1"
|
67
|
+
@mHeader[$TUNNEL_VERSION] = "3"
|
68
|
+
end
|
69
|
+
@mHeader[$USER_AGENT] = $SDK_UA_STR + @mOdpsConfig.userAgent
|
70
|
+
@mHeader[$AUTHORIZATION] = signAuthorization
|
71
|
+
case @mMethod
|
72
|
+
when "POST"
|
73
|
+
@mReq = ::Net::HTTP::Post.new(mUri.to_s, @mHeader)
|
74
|
+
@mReq.body = @mStream
|
75
|
+
when "GET"
|
76
|
+
@mReq = ::Net::HTTP::Get.new(mUri.to_s, @mHeader)
|
77
|
+
when "PUT"
|
78
|
+
@mReq = ::Net::HTTP::Put.new(mUri.to_s, @mHeader)
|
79
|
+
@mReq.body = @mStream
|
80
|
+
else
|
81
|
+
raise OdpsDatahubException.new($INVALID_ARGUMENT, "invalid method")
|
82
|
+
end
|
83
|
+
end
|
84
|
+
|
85
|
+
def getResponse()
|
86
|
+
res = Net::HTTP.start(@mUri.host, @mUri.port) {|http|
|
87
|
+
http.request(@mReq)
|
88
|
+
}
|
89
|
+
return res
|
90
|
+
end
|
91
|
+
|
92
|
+
def signAuthorization
|
93
|
+
prefix = "x-odps-"
|
94
|
+
stringToSign = @mMethod + "\n"
|
95
|
+
accessKey = @mOdpsConfig.accessKey
|
96
|
+
headerMapDown = Hash.new
|
97
|
+
|
98
|
+
@mHeader.each { |key , value|
|
99
|
+
keyDown = key.downcase
|
100
|
+
headerMapDown[keyDown] = value
|
101
|
+
}
|
102
|
+
headerArray = headerMapDown.sort
|
103
|
+
|
104
|
+
headerArray.each { |key , value|
|
105
|
+
if key.start_with?(prefix)
|
106
|
+
stringToSign << key << ":" << value
|
107
|
+
stringToSign << "\n"
|
108
|
+
elsif key == 'content-type' or key == 'content-md5' or key == 'date'
|
109
|
+
stringToSign << value
|
110
|
+
stringToSign << "\n"
|
111
|
+
end
|
112
|
+
}
|
113
|
+
|
114
|
+
signParam = ""
|
115
|
+
separater = '?'
|
116
|
+
paramArray = @mParam.sort
|
117
|
+
paramArray.each { |key , value|
|
118
|
+
if value != ""
|
119
|
+
signParam += separater + key.to_s + '=' + value.to_s
|
120
|
+
else
|
121
|
+
signParam += separater + key.to_s
|
122
|
+
end
|
123
|
+
separater = '&'
|
124
|
+
}
|
125
|
+
stringToSign += @mPath + signParam
|
126
|
+
#puts stringToSign
|
127
|
+
signedStr = "ODPS " + @mOdpsConfig.accessId + ":" + Base64.encode64("#{OpenSSL::HMAC.digest('sha1', accessKey, stringToSign)}").to_s
|
128
|
+
end
|
129
|
+
end
|
130
|
+
end
|
@@ -0,0 +1,73 @@
|
|
1
|
+
#
|
2
|
+
#Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
#or more contributor license agreements. See the NOTICE file
|
4
|
+
#distributed with this work for additional information
|
5
|
+
#regarding copyright ownership. The ASF licenses this file
|
6
|
+
#to you under the Apache License, Version 2.0 (the
|
7
|
+
#"License"); you may not use this file except in compliance
|
8
|
+
#with the License. You may obtain a copy of the License at
|
9
|
+
#
|
10
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
#
|
12
|
+
#Unless required by applicable law or agreed to in writing,
|
13
|
+
#software distributed under the License is distributed on an
|
14
|
+
#"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
#KIND, either express or implied. See the License for the
|
16
|
+
#specific language governing permissions and limitations
|
17
|
+
#under the License.
|
18
|
+
#
|
19
|
+
module OdpsDatahub
|
20
|
+
$SDK_UA_STR = "ODPS Ruby SDK v0.1"
|
21
|
+
$MAX_PACK_SIZE = 2048*1024
|
22
|
+
class HttpHeaders
|
23
|
+
$AUTHORIZATION = "Authorization"
|
24
|
+
$CACHE_CONTROL = "Cache-Control"
|
25
|
+
$CONTENT_DISPOSITION = "Content-Disposition"
|
26
|
+
$CONTENT_ENCODING = "Content-Encoding"
|
27
|
+
$CONTENT_LENGTH = "Content-Length"
|
28
|
+
$CONTENT_MD5 = "Content-MD5"
|
29
|
+
$CONTENT_TYPE = "Content-Type"
|
30
|
+
$DATE = "Date"
|
31
|
+
$ETAG = "ETag"
|
32
|
+
$EXPIRES = "Expires"
|
33
|
+
$HOST = "Host"
|
34
|
+
$LAST_MODIFIED = "Last-Modified"
|
35
|
+
$RANGE = "Range"
|
36
|
+
$LOCATION = "Location"
|
37
|
+
$TRANSFER_ENCODING = "Transfer-Encoding"
|
38
|
+
$CHUNKED = "chunked"
|
39
|
+
$ACCEPT_ENCODING = "Accept-Encoding"
|
40
|
+
$USER_AGENT = "User-Agent"
|
41
|
+
$TUNNEL_VERSION = "x-odps-tunnel-version"
|
42
|
+
$TUNNEL_STREAM_VERSION = "x-odps-tunnel-stream-version"
|
43
|
+
end
|
44
|
+
class HttpParam
|
45
|
+
$PARAM_RECORD_COUNT = "recordcount"
|
46
|
+
$PARAM_PACK_ID = "packid"
|
47
|
+
$PARAM_PACK_NUM = "packnum"
|
48
|
+
$PARAM_ITERATE_MODE = "iteratemode"
|
49
|
+
$PARAM_ITER_MODE_AT_PACKID = "AT_PACKID"
|
50
|
+
$PARAM_ITER_MODE_AFTER_PACKID = "AFTER_PACKID"
|
51
|
+
$PARAM_ITER_MODE_FIRST_PACK = "FIRST_PACK"
|
52
|
+
$PARAM_ITER_MODE_LAST_PACK = "LAST_PACK"
|
53
|
+
$PARAM_SHARD_NUMBER = "shardnumber"
|
54
|
+
$PARAM_SHARD_STATUS = "shardstatus"
|
55
|
+
$PARAM_PARTITION = "partition"
|
56
|
+
$PARAM_PARTITIONS = "partitions"
|
57
|
+
$PARAM_SEEK_TIME = "timestamp"
|
58
|
+
$PARAM_CURR_PROJECT = "curr_project"
|
59
|
+
$PARAM_TYPE = "type"
|
60
|
+
$PARAM_QUERY = "query"
|
61
|
+
$PARAM_EXPECT_MARKER = "expectmarker"
|
62
|
+
$PARAM_MARKER = "marker"
|
63
|
+
end
|
64
|
+
class PackType
|
65
|
+
@@FIRST_PACK_ID = "00000000000000000000000000000000"
|
66
|
+
end
|
67
|
+
class ReadMode
|
68
|
+
@@SEEK_BEGIN = 1
|
69
|
+
@@SEEK_END = 2
|
70
|
+
@@SEEK_CUR = 3
|
71
|
+
@@SEEK_NEXT = 4
|
72
|
+
end
|
73
|
+
end
|
@@ -0,0 +1,253 @@
|
|
1
|
+
#
|
2
|
+
#Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
#or more contributor license agreements. See the NOTICE file
|
4
|
+
#distributed with this work for additional information
|
5
|
+
#regarding copyright ownership. The ASF licenses this file
|
6
|
+
#to you under the Apache License, Version 2.0 (the
|
7
|
+
#"License"); you may not use this file except in compliance
|
8
|
+
#with the License. You may obtain a copy of the License at
|
9
|
+
#
|
10
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
#
|
12
|
+
#Unless required by applicable law or agreed to in writing,
|
13
|
+
#software distributed under the License is distributed on an
|
14
|
+
#"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
#KIND, either express or implied. See the License for the
|
16
|
+
#specific language governing permissions and limitations
|
17
|
+
#under the License.
|
18
|
+
#
|
19
|
+
require 'rexml/document'
|
20
|
+
require_relative 'xml_template'
|
21
|
+
require_relative 'odps_table_schema'
|
22
|
+
require_relative '../http/http_connection'
|
23
|
+
|
24
|
+
module OdpsDatahub
|
25
|
+
$STRING_MAX_LENTH = 8 * 1024 * 1024
|
26
|
+
$DATETIME_MAX_TICKS = 253402271999000
|
27
|
+
$DATETIME_MIN_TICKS = -62135798400000
|
28
|
+
$STRING_CHARSET = "UTF-8"
|
29
|
+
class OdpsTableRecord
|
30
|
+
attr_reader :mValues, :mSchema
|
31
|
+
|
32
|
+
def initialize(schema)
|
33
|
+
@mSchema = schema
|
34
|
+
@mValues = Array.new(@mSchema.getColumnCount)
|
35
|
+
end
|
36
|
+
|
37
|
+
def getColumnsCount
|
38
|
+
@mSchema.getColumnCount
|
39
|
+
end
|
40
|
+
|
41
|
+
def getTableSchema
|
42
|
+
@mSchema
|
43
|
+
end
|
44
|
+
|
45
|
+
def getValue(idx)
|
46
|
+
if idx < 0 or idx >= @mSchema.getColumnCount
|
47
|
+
raise "idx out of range"
|
48
|
+
end
|
49
|
+
@mValues.at(idx)
|
50
|
+
end
|
51
|
+
|
52
|
+
def setNullValue(idx)
|
53
|
+
setValue(idx, nil)
|
54
|
+
end
|
55
|
+
|
56
|
+
def setBigInt(idx, value)
|
57
|
+
if value.is_a?Integer
|
58
|
+
setValue(idx, value)
|
59
|
+
else
|
60
|
+
raise "value show be Integer"
|
61
|
+
end
|
62
|
+
end
|
63
|
+
|
64
|
+
def setDouble(idx, value)
|
65
|
+
if value.is_a?Float
|
66
|
+
setValue(idx, value)
|
67
|
+
else
|
68
|
+
raise "value show be Float"
|
69
|
+
end
|
70
|
+
end
|
71
|
+
|
72
|
+
def setBoolean(idx, value)
|
73
|
+
if value != false and value != true
|
74
|
+
raise "value must be bool"
|
75
|
+
end
|
76
|
+
setValue(idx, value)
|
77
|
+
end
|
78
|
+
|
79
|
+
def setDateTime(idx, value)
|
80
|
+
if value.is_a?Integer and value >= $DATETIME_MIN_TICKS and value <=> $DATETIME_MAX_TICKS
|
81
|
+
setValue(idx, value)
|
82
|
+
else
|
83
|
+
raise "DateTime out of range or value show be Integer"
|
84
|
+
end
|
85
|
+
end
|
86
|
+
|
87
|
+
def setString(idx, value)
|
88
|
+
if value.is_a?String and value.length < $STRING_MAX_LENTH
|
89
|
+
setValue(idx, value)
|
90
|
+
else
|
91
|
+
raise "value show be String and len < #$STRING_MAX_LENTH"
|
92
|
+
end
|
93
|
+
end
|
94
|
+
|
95
|
+
private
|
96
|
+
def setValue(idx, value)
|
97
|
+
if idx < 0 or idx >= @mSchema.getColumnCount
|
98
|
+
raise "idx out of range"
|
99
|
+
end
|
100
|
+
@mValues[idx] = value
|
101
|
+
end
|
102
|
+
end
|
103
|
+
|
104
|
+
class OdpsTable
|
105
|
+
def initialize(odpsConfig, projectName, tableName)
|
106
|
+
@mOdpsConfig = odpsConfig
|
107
|
+
@mProjectName = projectName
|
108
|
+
@mTableName = tableName
|
109
|
+
end
|
110
|
+
|
111
|
+
#get partitions and return an array like :[{"time"=>"2016", "place"=>"china2"},{"time"=>"2015", "place"=>"china"}]
|
112
|
+
def getPartitionList
|
113
|
+
partitionList = Array.new
|
114
|
+
url = "/projects/" + @mProjectName +"/tables/" + @mTableName
|
115
|
+
lastMarker = nil
|
116
|
+
isEnd = false
|
117
|
+
while !isEnd do
|
118
|
+
header = Hash.new
|
119
|
+
param = Hash.new
|
120
|
+
param[$PARAM_CURR_PROJECT] = @mProjectName
|
121
|
+
param[$PARAM_EXPECT_MARKER] = true
|
122
|
+
param[$PARAM_PARTITIONS] = ""
|
123
|
+
if lastMarker != nil
|
124
|
+
param[$PARAM_MARKER] = lastMarker
|
125
|
+
end
|
126
|
+
conn = HttpConnection.new(@mOdpsConfig, header, param, url, "GET", "", true)
|
127
|
+
res = conn.getResponse
|
128
|
+
if res.code != "200"
|
129
|
+
return partitionList
|
130
|
+
#raise OdpsDatahubException.new($INVALID_ARGUMENT, "This not a partitioned table")
|
131
|
+
end
|
132
|
+
|
133
|
+
doc = REXML::Document.new(res.body.to_s)
|
134
|
+
|
135
|
+
#parse partitions
|
136
|
+
partitionsXml = doc.root.get_elements("Partition")
|
137
|
+
partitionsXml.each { |partition|
|
138
|
+
partitionInfo = Hash.new
|
139
|
+
partition.elements.each { |column|
|
140
|
+
partitionInfo[column.attributes["Name"]] = column.attributes["Value"]
|
141
|
+
}
|
142
|
+
partitionList.push(partitionInfo)
|
143
|
+
}
|
144
|
+
|
145
|
+
#get marker
|
146
|
+
markerXml = doc.root.get_elements("Marker")
|
147
|
+
if markerXml[0].text == nil
|
148
|
+
isEnd = true
|
149
|
+
elsif
|
150
|
+
lastMarker = markerXml[0].text
|
151
|
+
end
|
152
|
+
end
|
153
|
+
return partitionList
|
154
|
+
end
|
155
|
+
|
156
|
+
#ptStr ex: 'dt=20150805,hh=08,mm=24'
|
157
|
+
#call add partiton if not exsits
|
158
|
+
def addPartition(ptStr)
|
159
|
+
pts_array = ptStr.split(',')
|
160
|
+
sqlstr = "ALTER TABLE " + @mProjectName + "." + @mTableName
|
161
|
+
sqlstr = sqlstr + " ADD IF NOT EXISTS" + " PARTITION ("
|
162
|
+
pts_array.each { |pt|
|
163
|
+
ptkv = pt.split('=')
|
164
|
+
if ptkv.size != 2
|
165
|
+
raise "invalid partition spec" + pt
|
166
|
+
end
|
167
|
+
sqlstr += ptkv[0] + '=' + "'" + ptkv[1] + "'" + ','
|
168
|
+
}
|
169
|
+
sqlstr = sqlstr[0..-2] + ");"
|
170
|
+
taskName = "SQLAddPartitionTask"
|
171
|
+
runSQL(taskName, sqlstr)
|
172
|
+
end
|
173
|
+
|
174
|
+
def runSQL(taskName, sqlstring)
|
175
|
+
task_xml = XmlTemplate.getTaskXml(taskName, sqlstring)
|
176
|
+
|
177
|
+
job_xml = genJobXml('arbitriary_job', '9', "", task_xml)
|
178
|
+
headers = Hash.new
|
179
|
+
headers['Content-Type'] = 'application/xml'
|
180
|
+
headers['Content-MD5'] = Digest::MD5.hexdigest(job_xml)
|
181
|
+
headers['Content-Length'] = job_xml.size.to_s
|
182
|
+
|
183
|
+
params = Hash.new
|
184
|
+
|
185
|
+
url = "/projects/" + @mProjectName +"/instances"
|
186
|
+
conn = HttpConnection.new(@mOdpsConfig, headers, params, url, 'POST', job_xml, true)
|
187
|
+
|
188
|
+
res = conn.getResponse
|
189
|
+
if res.code != '200'
|
190
|
+
raise "Add partition failed with error" + res.code.to_s
|
191
|
+
end
|
192
|
+
|
193
|
+
if res.to_hash['Content-Length'] != "0" and not res.body.to_s.include?"Instance"
|
194
|
+
raise res.body
|
195
|
+
end
|
196
|
+
|
197
|
+
waitForSQLComplete(res)
|
198
|
+
end
|
199
|
+
|
200
|
+
#TODO support mulit task
|
201
|
+
def genJobXml(name, priority, comment, taskStr, runMode='sequence')
|
202
|
+
job_xml = XmlTemplate.getJobXml(name, priority, comment, taskStr, runMode)
|
203
|
+
return job_xml
|
204
|
+
end
|
205
|
+
|
206
|
+
def waitForSQLComplete(res)
|
207
|
+
ret_headers = res.to_hash
|
208
|
+
instanceurl = "/projects/" + @mProjectName +"/instances" + "/" + ret_headers['location'][0].split('/')[-1]
|
209
|
+
|
210
|
+
headers = Hash.new
|
211
|
+
params = Hash.new
|
212
|
+
params['taskstatus'] = ""
|
213
|
+
res = nil
|
214
|
+
|
215
|
+
while true
|
216
|
+
conn = HttpConnection.new(@mOdpsConfig, headers, params, instanceurl, 'GET', "", true)
|
217
|
+
res = conn.getResponse
|
218
|
+
doc = REXML::Document.new(res.body.to_s)
|
219
|
+
insStatus = doc.root.elements["Status"].text
|
220
|
+
if insStatus == 'Terminated'
|
221
|
+
break;
|
222
|
+
elsif insStatus == 'Running' or insStatus == 'Suspended'
|
223
|
+
sleep(5)
|
224
|
+
end
|
225
|
+
end
|
226
|
+
|
227
|
+
doc.root.elements.each('Tasks/Task') { |e|
|
228
|
+
status = e.elements['Status'].text
|
229
|
+
name = e.elements['Name'].text
|
230
|
+
if status.to_s != 'Success'
|
231
|
+
raise getTaskResult(instanceurl, name.to_s)
|
232
|
+
end
|
233
|
+
}
|
234
|
+
end
|
235
|
+
|
236
|
+
def getTaskResult(instanceurl, name)
|
237
|
+
headers = Hash.new
|
238
|
+
params = Hash.new
|
239
|
+
params['result'] = ""
|
240
|
+
res = nil
|
241
|
+
|
242
|
+
conn = HttpConnection.new(@mOdpsConfig, headers, params, instanceurl, 'GET', "", true)
|
243
|
+
res = conn.getResponse
|
244
|
+
doc = REXML::Document.new(res.body.to_s)
|
245
|
+
doc.root.elements.each('Tasks/Task') { |e|
|
246
|
+
taskname = e.elements['Name'].text
|
247
|
+
if taskname == name.to_s
|
248
|
+
return e.elements['Result'].cdatas().to_s
|
249
|
+
end
|
250
|
+
}
|
251
|
+
end
|
252
|
+
end
|
253
|
+
end
|