【发布时间】：2018-07-16 08:09:15
【问题描述】：
引发了不起眼的异常：
Traceback (most recent call last):
File "run1.py", line 44, in <module>
bl = list(bl_trees)
File "/home/zal/.local/lib/python3.5/site-packages/multiprocess/pool.py", line 731, in next
raise value
IndexError: list index out of range
代码的简化版本是：
i = 0
for workload in workloads:
print(workload)
for M, dM in [(1000, 50)]:
print(M, dM)
try:
print("Baseline.")
bl_trees = LSMulator.cache_vs_bloom_vs_buf_threaded(workload, M, dM=dM, balloc=baseline_assignment, verbose=True)
bl = list(bl_trees)
for t in bl:
t.clear_data()
with open('./results-{}-{}-{}-{}.pkl'.format(str(workload), M, dM, "bl"), 'wb') as f:
pickle.dump((str(workload), M, dM, bl), f)
except Exception as ex:
traceback.print_exc()
try:
print("Monkey.")
mk_trees = LSMulator.cache_vs_bloom_vs_buf_threaded(workload, M, dM=dM, balloc=monkey_assignment, verbose=True)
mk = list(mk_trees)
for t in mk:
t.clear_data()
with open('./results-{}-{}-{}-{}.pkl'.format(str(workload), M, dM, "mk"), 'wb') as f:
pickle.dump((str(workload), M, dM, mk), f)
except Exception as ex:
traceback.print_exc()
i += 1
这是完整的代码：
from lsmulator import *
from workloads import *
from bloom_assignments import *
from workloads import *
import scipy.stats
import numpy as np
import dill as pickle
import traceback
n = 50000
k = 5000
workloads = [
UniformWorkload(n_queries=n, k_classes=k),
RoundRobinWorkload(n_queries=n, k_classes=k),
ZipfWorkload(n_queries=n, zipf_param=1.1),
ZipfWorkload(n_queries=n, zipf_param=1.5),
ZipfWorkload(n_queries=n, zipf_param=1.8),
EightyTwentyWorkload(n_queries=n, k_classes=k),
EightyTwentyWorkload(n_queries=n, k_classes=k, update_fraction=0.25),
EightyTwentyWorkload(n_queries=n, k_classes=k, update_fraction=0.65),
MultinomialWorkload(n_queries=n, k_classes=k),
MultinomialWorkload(n_queries=n, k_classes=k, dist=scipy.stats.gamma(2)),
MultinomialWorkload(n_queries=n, k_classes=k, dist=scipy.stats.expon(1)),
DiscoverDecayWorkload(n_queries=n),
DiscoverDecayWorkload(n_queries=n, lookups=scipy.stats.poisson(8)),
DiscoverDecayWorkload(n_queries=n, decay_rate=scipy.stats.beta(2, 1)),
DiscoverDecayWorkload(n_queries=n, updates=scipy.stats.poisson(0), decay_rate=scipy.stats.beta(2, 1)),
DiscoverDecayWorkload(n_queries=n, updates=scipy.stats.poisson(16), decay_rate=scipy.stats.beta(2, 1)),
PeriodicDecayWorkload(n_queries=n, period=100),
PeriodicDecayWorkload(n_queries=n, period=1000),
PeriodicDecayWorkload(n_queries=n, period=1000, decay_rate=scipy.stats.beta(2, 1))
]
i = 0
for workload in workloads:
print(workload)
for M, dM in [(1000, 50)]:
print(M, dM)
try:
print("Baseline.")
bl_trees = LSMulator.cache_vs_bloom_vs_buf_threaded(workload, M, dM=dM, balloc=baseline_assignment, verbose=True)
bl = list(bl_trees)
for t in bl:
t.clear_data()
with open('./results-{}-{}-{}-{}.pkl'.format(str(workload), M, dM, "bl"), 'wb') as f:
pickle.dump((str(workload), M, dM, bl), f)
except Exception as ex:
traceback.print_exc()
try:
print("Monkey.")
mk_trees = LSMulator.cache_vs_bloom_vs_buf_threaded(workload, M, dM=dM, balloc=monkey_assignment, verbose=True)
mk = list(mk_trees)
for t in mk:
t.clear_data()
with open('./results-{}-{}-{}-{}.pkl'.format(str(workload), M, dM, "mk"), 'wb') as f:
pickle.dump((str(workload), M, dM, mk), f)
except Exception as ex:
traceback.print_exc()
i += 1
【问题讨论】：
-
很遗憾，您的代码不是 Minimal, Complete, and Verifiable example，我们无法重现该错误。如果您访问超出其长度的列表，则会发生这种错误：
c= [1,2,3]&&print(c[8])。请提供 MVCE 或使用 How to debug small programs (#1) 来查找您的错误。 -
你能用
bl_trees = LSMulator...和bl = list(bl_trees)这两行复制它吗？如果有，bl_trees是什么？ -
这取决于
LSMulator.cache_vs_bloom_vs_buf_threaded返回的内容。LSMulator是你自己的代码还是库？ -
LSMulator是我自己的代码，代码链接github.com/asross/cs265/blob/master/simulations/lsmulator.py