From 029aef2ee231abe44b4300e16801ca4941481ef1 Mon Sep 17 00:00:00 2001
From: Bert Palm <bert.palm@ufz.de>
Date: Fri, 21 Feb 2020 13:45:17 +0100
Subject: [PATCH] fixedes
---
dios/dios.py | 6 +++++-
dios/locator.py | 5 +++--
profiling/performance.py | 21 ++++++++++-----------
3 files changed, 18 insertions(+), 14 deletions(-)
diff --git a/dios/dios.py b/dios/dios.py
index b83358e..4c7d308 100644
--- a/dios/dios.py
+++ b/dios/dios.py
@@ -405,7 +405,11 @@ class DictOfSeries:
def squeeze(self, axis=None):
if axis in [0, 'index']:
raise NotImplementedError
- elif axis in [1, 'columns']:
+
+ if len(self) > 1:
+ return self
+
+ if axis in [1, 'columns']:
return self._data.squeeze()
elif axis is None:
return self._data.squeeze().squeeze()
diff --git a/dios/locator.py b/dios/locator.py
index ac7cbdc..cb2f8b4 100644
--- a/dios/locator.py
+++ b/dios/locator.py
@@ -108,13 +108,14 @@ class _iLocIndexer(_Indexer):
# we do use loc instead of iloc as we get real keys.
# this works, because keys keep sorted if they come
# from an series-index (doesn't work with df-index)
- ser = self._data[key]
+ ser = self._data.loc[key]
if ixalign:
ix = ser.index.intersection(ix.index)
if isinstance(right, pd.Series):
left = ser[ix]
right, ix = align_index_by_policy(left, right)
ser.iloc[ix] = right
+ self._data.loc[key] = ser
def _get_item(self, key, ix, ixalign=False):
ser = self._data.loc[key]
@@ -148,7 +149,7 @@ def _unpack_value(keys, ix, val):
# prepare value
val = list(val) if is_iterator(val) else val
- val = val.squeeze() if is_dios_like(val) else val
+ val = val.squeeze(axis=1) if is_dios_like(val) else val
dioslike, nlistlike = is_dios_like(val), is_nested_list_like(val)
# check value
diff --git a/profiling/performance.py b/profiling/performance.py
index 1be82e8..de2e58a 100644
--- a/profiling/performance.py
+++ b/profiling/performance.py
@@ -63,12 +63,12 @@ def b_timings(df, t0, t1, v1, v2):
def dios_timings(dios, t0, t1, v1, v2):
_t0 = time.time()
- a = dios[t0:t1, :]
+ a = dios.loc[t0:t1, :]
_t1 = time.time()
- b = dios[:, v1]
+ b = dios.loc[:, v1]
_t2 = time.time()
if profile_assignment:
- dios[t0:t1, v1] = dios[t0:t1, v1] * 1111
+ dios.loc[t0:t1, v1] = dios.loc[t0:t1, v1] * 1111
_t3 = time.time()
timingsdf.at[rows, ('ts', 'dios')] += _t1 - _t0
@@ -76,7 +76,6 @@ def dios_timings(dios, t0, t1, v1, v2):
timingsdf.at[rows, ('ass', 'dios')] += _t3 - _t2
return a, b, dios
-
def gen_random_timestamps(m, M):
r = (M - m) * (np.random.randint(10,90) + np.random.random()) * 0.01
a , b = m + r, M - r
@@ -106,22 +105,22 @@ if __name__ == '__main__':
# max increase of of rows
# 1 = 10 # 2 = 100 # .... # 5 = 100'000
- iterations = 4
- runs = 10
+ iterations = 5
+ runs = 1
cols = 10
profile_assignment = True
# which to calc and plot
- use_df = True
- use_a = False
- use_b = False
+ use_df = False
+ use_a = True
+ use_b = True
use_dios = True
# plot options
- normalize_to_df = False
+ normalize_to_df = True
plot_xlog = True
- plot_ylog = False
+ plot_ylog = True
# ########################
--
GitLab