|
| 1 | +from datetime import datetime |
| 2 | + |
| 3 | +import numpy as np |
| 4 | +import pandas as pd |
| 5 | +from pandas.util import testing as tm |
| 6 | +from pandas.compat import lrange |
| 7 | +from pandas import (PeriodIndex, Series, DatetimeIndex, |
| 8 | + period_range, Period, tslib, _np_version_under1p9) |
| 9 | + |
| 10 | + |
| 11 | +class TestGetItem(tm.TestCase): |
| 12 | + |
| 13 | + def setUp(self): |
| 14 | + pass |
| 15 | + |
| 16 | + def test_getitem(self): |
| 17 | + idx1 = pd.period_range('2011-01-01', '2011-01-31', freq='D', |
| 18 | + name='idx') |
| 19 | + |
| 20 | + for idx in [idx1]: |
| 21 | + result = idx[0] |
| 22 | + self.assertEqual(result, pd.Period('2011-01-01', freq='D')) |
| 23 | + |
| 24 | + result = idx[-1] |
| 25 | + self.assertEqual(result, pd.Period('2011-01-31', freq='D')) |
| 26 | + |
| 27 | + result = idx[0:5] |
| 28 | + expected = pd.period_range('2011-01-01', '2011-01-05', freq='D', |
| 29 | + name='idx') |
| 30 | + self.assert_index_equal(result, expected) |
| 31 | + self.assertEqual(result.freq, expected.freq) |
| 32 | + self.assertEqual(result.freq, 'D') |
| 33 | + |
| 34 | + result = idx[0:10:2] |
| 35 | + expected = pd.PeriodIndex(['2011-01-01', '2011-01-03', |
| 36 | + '2011-01-05', |
| 37 | + '2011-01-07', '2011-01-09'], |
| 38 | + freq='D', name='idx') |
| 39 | + self.assert_index_equal(result, expected) |
| 40 | + self.assertEqual(result.freq, expected.freq) |
| 41 | + self.assertEqual(result.freq, 'D') |
| 42 | + |
| 43 | + result = idx[-20:-5:3] |
| 44 | + expected = pd.PeriodIndex(['2011-01-12', '2011-01-15', |
| 45 | + '2011-01-18', |
| 46 | + '2011-01-21', '2011-01-24'], |
| 47 | + freq='D', name='idx') |
| 48 | + self.assert_index_equal(result, expected) |
| 49 | + self.assertEqual(result.freq, expected.freq) |
| 50 | + self.assertEqual(result.freq, 'D') |
| 51 | + |
| 52 | + result = idx[4::-1] |
| 53 | + expected = PeriodIndex(['2011-01-05', '2011-01-04', '2011-01-03', |
| 54 | + '2011-01-02', '2011-01-01'], |
| 55 | + freq='D', name='idx') |
| 56 | + self.assert_index_equal(result, expected) |
| 57 | + self.assertEqual(result.freq, expected.freq) |
| 58 | + self.assertEqual(result.freq, 'D') |
| 59 | + |
| 60 | + def test_getitem_index(self): |
| 61 | + idx = period_range('2007-01', periods=10, freq='M', name='x') |
| 62 | + |
| 63 | + result = idx[[1, 3, 5]] |
| 64 | + exp = pd.PeriodIndex(['2007-02', '2007-04', '2007-06'], |
| 65 | + freq='M', name='x') |
| 66 | + tm.assert_index_equal(result, exp) |
| 67 | + |
| 68 | + result = idx[[True, True, False, False, False, |
| 69 | + True, True, False, False, False]] |
| 70 | + exp = pd.PeriodIndex(['2007-01', '2007-02', '2007-06', '2007-07'], |
| 71 | + freq='M', name='x') |
| 72 | + tm.assert_index_equal(result, exp) |
| 73 | + |
| 74 | + def test_getitem_partial(self): |
| 75 | + rng = period_range('2007-01', periods=50, freq='M') |
| 76 | + ts = Series(np.random.randn(len(rng)), rng) |
| 77 | + |
| 78 | + self.assertRaises(KeyError, ts.__getitem__, '2006') |
| 79 | + |
| 80 | + result = ts['2008'] |
| 81 | + self.assertTrue((result.index.year == 2008).all()) |
| 82 | + |
| 83 | + result = ts['2008':'2009'] |
| 84 | + self.assertEqual(len(result), 24) |
| 85 | + |
| 86 | + result = ts['2008-1':'2009-12'] |
| 87 | + self.assertEqual(len(result), 24) |
| 88 | + |
| 89 | + result = ts['2008Q1':'2009Q4'] |
| 90 | + self.assertEqual(len(result), 24) |
| 91 | + |
| 92 | + result = ts[:'2009'] |
| 93 | + self.assertEqual(len(result), 36) |
| 94 | + |
| 95 | + result = ts['2009':] |
| 96 | + self.assertEqual(len(result), 50 - 24) |
| 97 | + |
| 98 | + exp = result |
| 99 | + result = ts[24:] |
| 100 | + tm.assert_series_equal(exp, result) |
| 101 | + |
| 102 | + ts = ts[10:].append(ts[10:]) |
| 103 | + self.assertRaisesRegexp(KeyError, |
| 104 | + "left slice bound for non-unique " |
| 105 | + "label: '2008'", |
| 106 | + ts.__getitem__, slice('2008', '2009')) |
| 107 | + |
| 108 | + def test_getitem_datetime(self): |
| 109 | + rng = period_range(start='2012-01-01', periods=10, freq='W-MON') |
| 110 | + ts = Series(lrange(len(rng)), index=rng) |
| 111 | + |
| 112 | + dt1 = datetime(2011, 10, 2) |
| 113 | + dt4 = datetime(2012, 4, 20) |
| 114 | + |
| 115 | + rs = ts[dt1:dt4] |
| 116 | + tm.assert_series_equal(rs, ts) |
| 117 | + |
| 118 | + def test_getitem_nat(self): |
| 119 | + idx = pd.PeriodIndex(['2011-01', 'NaT', '2011-02'], freq='M') |
| 120 | + self.assertEqual(idx[0], pd.Period('2011-01', freq='M')) |
| 121 | + self.assertIs(idx[1], tslib.NaT) |
| 122 | + |
| 123 | + s = pd.Series([0, 1, 2], index=idx) |
| 124 | + self.assertEqual(s[pd.NaT], 1) |
| 125 | + |
| 126 | + s = pd.Series(idx, index=idx) |
| 127 | + self.assertEqual(s[pd.Period('2011-01', freq='M')], |
| 128 | + pd.Period('2011-01', freq='M')) |
| 129 | + self.assertIs(s[pd.NaT], tslib.NaT) |
| 130 | + |
| 131 | + def test_getitem_list_periods(self): |
| 132 | + # GH 7710 |
| 133 | + rng = period_range(start='2012-01-01', periods=10, freq='D') |
| 134 | + ts = Series(lrange(len(rng)), index=rng) |
| 135 | + exp = ts.iloc[[1]] |
| 136 | + tm.assert_series_equal(ts[[Period('2012-01-02', freq='D')]], exp) |
| 137 | + |
| 138 | + def test_getitem_seconds(self): |
| 139 | + # GH 6716 |
| 140 | + didx = DatetimeIndex(start='2013/01/01 09:00:00', freq='S', |
| 141 | + periods=4000) |
| 142 | + pidx = PeriodIndex(start='2013/01/01 09:00:00', freq='S', periods=4000) |
| 143 | + |
| 144 | + for idx in [didx, pidx]: |
| 145 | + # getitem against index should raise ValueError |
| 146 | + values = ['2014', '2013/02', '2013/01/02', '2013/02/01 9H', |
| 147 | + '2013/02/01 09:00'] |
| 148 | + for v in values: |
| 149 | + if _np_version_under1p9: |
| 150 | + with tm.assertRaises(ValueError): |
| 151 | + idx[v] |
| 152 | + else: |
| 153 | + # GH7116 |
| 154 | + # these show deprecations as we are trying |
| 155 | + # to slice with non-integer indexers |
| 156 | + # with tm.assertRaises(IndexError): |
| 157 | + # idx[v] |
| 158 | + continue |
| 159 | + |
| 160 | + s = Series(np.random.rand(len(idx)), index=idx) |
| 161 | + tm.assert_series_equal(s['2013/01/01 10:00'], s[3600:3660]) |
| 162 | + tm.assert_series_equal(s['2013/01/01 9H'], s[:3600]) |
| 163 | + for d in ['2013/01/01', '2013/01', '2013']: |
| 164 | + tm.assert_series_equal(s[d], s) |
| 165 | + |
| 166 | + def test_getitem_day(self): |
| 167 | + # GH 6716 |
| 168 | + # Confirm DatetimeIndex and PeriodIndex works identically |
| 169 | + didx = DatetimeIndex(start='2013/01/01', freq='D', periods=400) |
| 170 | + pidx = PeriodIndex(start='2013/01/01', freq='D', periods=400) |
| 171 | + |
| 172 | + for idx in [didx, pidx]: |
| 173 | + # getitem against index should raise ValueError |
| 174 | + values = ['2014', '2013/02', '2013/01/02', '2013/02/01 9H', |
| 175 | + '2013/02/01 09:00'] |
| 176 | + for v in values: |
| 177 | + |
| 178 | + if _np_version_under1p9: |
| 179 | + with tm.assertRaises(ValueError): |
| 180 | + idx[v] |
| 181 | + else: |
| 182 | + # GH7116 |
| 183 | + # these show deprecations as we are trying |
| 184 | + # to slice with non-integer indexers |
| 185 | + # with tm.assertRaises(IndexError): |
| 186 | + # idx[v] |
| 187 | + continue |
| 188 | + |
| 189 | + s = Series(np.random.rand(len(idx)), index=idx) |
| 190 | + tm.assert_series_equal(s['2013/01'], s[0:31]) |
| 191 | + tm.assert_series_equal(s['2013/02'], s[31:59]) |
| 192 | + tm.assert_series_equal(s['2014'], s[365:]) |
| 193 | + |
| 194 | + invalid = ['2013/02/01 9H', '2013/02/01 09:00'] |
| 195 | + for v in invalid: |
| 196 | + with tm.assertRaises(KeyError): |
| 197 | + s[v] |
| 198 | + |
| 199 | + |
| 200 | +class TestIndexing(tm.TestCase): |
| 201 | + |
| 202 | + def test_get_loc_msg(self): |
| 203 | + idx = period_range('2000-1-1', freq='A', periods=10) |
| 204 | + bad_period = Period('2012', 'A') |
| 205 | + self.assertRaises(KeyError, idx.get_loc, bad_period) |
| 206 | + |
| 207 | + try: |
| 208 | + idx.get_loc(bad_period) |
| 209 | + except KeyError as inst: |
| 210 | + self.assertEqual(inst.args[0], bad_period) |
| 211 | + |
| 212 | + def test_get_loc_nat(self): |
| 213 | + didx = DatetimeIndex(['2011-01-01', 'NaT', '2011-01-03']) |
| 214 | + pidx = PeriodIndex(['2011-01-01', 'NaT', '2011-01-03'], freq='M') |
| 215 | + |
| 216 | + # check DatetimeIndex compat |
| 217 | + for idx in [didx, pidx]: |
| 218 | + self.assertEqual(idx.get_loc(pd.NaT), 1) |
| 219 | + self.assertEqual(idx.get_loc(None), 1) |
| 220 | + self.assertEqual(idx.get_loc(float('nan')), 1) |
| 221 | + self.assertEqual(idx.get_loc(np.nan), 1) |
| 222 | + |
| 223 | + def test_take(self): |
| 224 | + # GH 10295 |
| 225 | + idx1 = pd.period_range('2011-01-01', '2011-01-31', freq='D', |
| 226 | + name='idx') |
| 227 | + |
| 228 | + for idx in [idx1]: |
| 229 | + result = idx.take([0]) |
| 230 | + self.assertEqual(result, pd.Period('2011-01-01', freq='D')) |
| 231 | + |
| 232 | + result = idx.take([5]) |
| 233 | + self.assertEqual(result, pd.Period('2011-01-06', freq='D')) |
| 234 | + |
| 235 | + result = idx.take([0, 1, 2]) |
| 236 | + expected = pd.period_range('2011-01-01', '2011-01-03', freq='D', |
| 237 | + name='idx') |
| 238 | + self.assert_index_equal(result, expected) |
| 239 | + self.assertEqual(result.freq, 'D') |
| 240 | + self.assertEqual(result.freq, expected.freq) |
| 241 | + |
| 242 | + result = idx.take([0, 2, 4]) |
| 243 | + expected = pd.PeriodIndex(['2011-01-01', '2011-01-03', |
| 244 | + '2011-01-05'], freq='D', name='idx') |
| 245 | + self.assert_index_equal(result, expected) |
| 246 | + self.assertEqual(result.freq, expected.freq) |
| 247 | + self.assertEqual(result.freq, 'D') |
| 248 | + |
| 249 | + result = idx.take([7, 4, 1]) |
| 250 | + expected = pd.PeriodIndex(['2011-01-08', '2011-01-05', |
| 251 | + '2011-01-02'], |
| 252 | + freq='D', name='idx') |
| 253 | + self.assert_index_equal(result, expected) |
| 254 | + self.assertEqual(result.freq, expected.freq) |
| 255 | + self.assertEqual(result.freq, 'D') |
| 256 | + |
| 257 | + result = idx.take([3, 2, 5]) |
| 258 | + expected = PeriodIndex(['2011-01-04', '2011-01-03', '2011-01-06'], |
| 259 | + freq='D', name='idx') |
| 260 | + self.assert_index_equal(result, expected) |
| 261 | + self.assertEqual(result.freq, expected.freq) |
| 262 | + self.assertEqual(result.freq, 'D') |
| 263 | + |
| 264 | + result = idx.take([-3, 2, 5]) |
| 265 | + expected = PeriodIndex(['2011-01-29', '2011-01-03', '2011-01-06'], |
| 266 | + freq='D', name='idx') |
| 267 | + self.assert_index_equal(result, expected) |
| 268 | + self.assertEqual(result.freq, expected.freq) |
| 269 | + self.assertEqual(result.freq, 'D') |
| 270 | + |
| 271 | + def test_take_misc(self): |
| 272 | + index = PeriodIndex(start='1/1/10', end='12/31/12', freq='D', |
| 273 | + name='idx') |
| 274 | + expected = PeriodIndex([datetime(2010, 1, 6), datetime(2010, 1, 7), |
| 275 | + datetime(2010, 1, 9), datetime(2010, 1, 13)], |
| 276 | + freq='D', name='idx') |
| 277 | + |
| 278 | + taken1 = index.take([5, 6, 8, 12]) |
| 279 | + taken2 = index[[5, 6, 8, 12]] |
| 280 | + |
| 281 | + for taken in [taken1, taken2]: |
| 282 | + tm.assert_index_equal(taken, expected) |
| 283 | + tm.assertIsInstance(taken, PeriodIndex) |
| 284 | + self.assertEqual(taken.freq, index.freq) |
| 285 | + self.assertEqual(taken.name, expected.name) |
| 286 | + |
| 287 | + def test_take_fill_value(self): |
| 288 | + # GH 12631 |
| 289 | + idx = pd.PeriodIndex(['2011-01-01', '2011-02-01', '2011-03-01'], |
| 290 | + name='xxx', freq='D') |
| 291 | + result = idx.take(np.array([1, 0, -1])) |
| 292 | + expected = pd.PeriodIndex(['2011-02-01', '2011-01-01', '2011-03-01'], |
| 293 | + name='xxx', freq='D') |
| 294 | + tm.assert_index_equal(result, expected) |
| 295 | + |
| 296 | + # fill_value |
| 297 | + result = idx.take(np.array([1, 0, -1]), fill_value=True) |
| 298 | + expected = pd.PeriodIndex(['2011-02-01', '2011-01-01', 'NaT'], |
| 299 | + name='xxx', freq='D') |
| 300 | + tm.assert_index_equal(result, expected) |
| 301 | + |
| 302 | + # allow_fill=False |
| 303 | + result = idx.take(np.array([1, 0, -1]), allow_fill=False, |
| 304 | + fill_value=True) |
| 305 | + expected = pd.PeriodIndex(['2011-02-01', '2011-01-01', '2011-03-01'], |
| 306 | + name='xxx', freq='D') |
| 307 | + tm.assert_index_equal(result, expected) |
| 308 | + |
| 309 | + msg = ('When allow_fill=True and fill_value is not None, ' |
| 310 | + 'all indices must be >= -1') |
| 311 | + with tm.assertRaisesRegexp(ValueError, msg): |
| 312 | + idx.take(np.array([1, 0, -2]), fill_value=True) |
| 313 | + with tm.assertRaisesRegexp(ValueError, msg): |
| 314 | + idx.take(np.array([1, 0, -5]), fill_value=True) |
| 315 | + |
| 316 | + with tm.assertRaises(IndexError): |
| 317 | + idx.take(np.array([1, -5])) |
0 commit comments