1. Adam Kawa

2. A deal was made!

6. Martin will invite Adam
and Timbuktu, my favourite Swedish artist,
for a beer or coke or whatever to drink *
by Martin

7. -

8. Question

9. Question
Answers

10. Data will tell the truth!

11. -
-
-
-
-

12. Why?
by Adam

13. -
-
-
-
-
-

14. Introduction

16. … …
… …
… …
… …
… …

17. …
…
…
…

18. -

19. ✓
✓
✓
✗
✗

20. HiveQL

30. A line where
I may have a bug ? !

31. HiveQL

32. Verbose
and
complex
Java
code

33. -

34. -
-

35. -

36. -
-

37. -
-
-

38. -
-
-
-

39. -
-
-
-
-

41. For Each Line
-

42. For Each Line
-
-

43. track.txt

44. user.txt
track.txt

45. stream.txt
user.txt
track.txt

46. expected.txt
stream.txt

49. …
…

50. …
…
…

51. …
…
…

52. Bee test
Be happy !

53. HiveQL

54. -
-
-

55. ✗
-
✗
-

56. … … …
✗

57. -
-
-

58. Threshold

59. ✓
-
Threshold

60. ✗
Threshold

61. ✗
Try and see
-

62. -
?

63. HiveQL

64. -

65. -
-
-
-

66. 2 MapReduce
job in total

67. Runs many Map
joins in a Map-Only
job
[HIVE-3784]

68. -
-
-

69. -
-
-
-

70. -
-
-
-
-
-
-

71. -
-
-
-
-
-
-

72. -
-
-

73. Runs as a single
MR job
[HIVE-3952]

74. 2 MapReduce
job in total

75. HiveQL

77. ✗
✓
-
-

78. ✓
✗
-

80. -

81. -
My query generates small amount of
intermediate data
-

82. ✓
✗

85. -

86. -
-
-
-
-
-

87. -
-
-
-
-
-
-
-

88. 2 months of data
50 min 2 sec
10th place
?

89. Changes are needed!

90. File Format

91. -
-

92. ✓
-
-

93. ✗
✓
-

94. 16x

95. 3.5x

96. 32x

97. -

98. Computation

99. -
-
-

100. 1.4x
2.4x

101. 8x

102. ✓
-
-
-

103. ✓
-
-

104. ✓
-
-

105. Time

106. The more congested
queue/cluster, the bigger
benefits of reusing
Time

107. No scheduling overhead to
run new Reduce task
Time

108. Time
Thinner tasks allows to
avoid stragglers

109. Finished within 1,5 sec.
Warm !

111. -
-
-
-

112. -

113. ✓
✓

114. -
-
-
-

115. Feature

116. ✓
-
✓
✓
-
✓

117. 1.4x

118. ✗
✓
✓
✗
✓

119. -
-

120. Feature

121. ✓
✓
-
-
-
✓

123. 14 months of data
10 min 11 sec
?

124. Results

127. That’s all !

128. -
-
-
-
-

129. -
-
-