mirror of
https://github.com/meteoinfo/MeteoInfo.git
synced 2025-12-08 20:36:05 +00:00
add reindex method in DataFrame class
This commit is contained in:
wyq
parent
ef5cae235b
commit
9e55114aa8
@ -2024,6 +2024,115 @@ public class DataFrame implements Iterable {
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Reindex by row and column ranges
|
||||
*
|
||||
* @param rowKeys Row keys
|
||||
* @param rowRange Row range
|
||||
* @param colRange Column indices
|
||||
* @param colNames Column names
|
||||
* @return Reindex data frame or series
|
||||
* @throws org.meteoinfo.ndarray.InvalidRangeException
|
||||
*/
|
||||
public Object reIndex(List rowKeys, List<Integer> rowRange, List<Integer> colRange,
|
||||
List<String> colNames) throws InvalidRangeException {
|
||||
ColumnIndex cols = new ColumnIndex();
|
||||
int idx = 0;
|
||||
for (int i : colRange) {
|
||||
if (i >= 0) {
|
||||
cols.add(this.columns.get(i));
|
||||
} else {
|
||||
cols.add(new Column(colNames.get(idx), DataType.FLOAT));
|
||||
}
|
||||
idx += 1;
|
||||
}
|
||||
|
||||
Object r;
|
||||
if (this.array2D) {
|
||||
int n = ((Array) data).getShape()[1];
|
||||
int rn = rowRange.size();
|
||||
int cn = colRange.size();
|
||||
DataType dtype = ((Array) data).getDataType();
|
||||
r = Array.factory(dtype, new int[]{rn, cn});
|
||||
String format = this.columns.get(0).getFormat();
|
||||
Object v = DataTypeUtil.convertTo(null, dtype, format);
|
||||
idx = 0;
|
||||
int jj = 0;
|
||||
for (int j : colRange) {
|
||||
if (j < 0) {
|
||||
int ii = 0;
|
||||
for (int i : rowRange) {
|
||||
idx = ii * cn + jj;
|
||||
((Array) r).setObject(idx, v);
|
||||
ii += 1;
|
||||
}
|
||||
} else {
|
||||
int ii = 0;
|
||||
for (int i : rowRange) {
|
||||
idx = ii * cn + jj;
|
||||
if (i < 0) {
|
||||
((Array) r).setObject(idx, v);
|
||||
} else {
|
||||
((Array) r).setObject(idx, ((Array) data).getObject(i * n + j));
|
||||
}
|
||||
ii += 1;
|
||||
}
|
||||
}
|
||||
jj += 1;
|
||||
}
|
||||
} else {
|
||||
r = new ArrayList<>();
|
||||
int rn = rowRange.size();
|
||||
for (int j : colRange) {
|
||||
Array rr;
|
||||
idx = 0;
|
||||
if (j < 0) {
|
||||
rr = Array.factory(DataType.FLOAT, new int[]{rn});
|
||||
for (int i : rowRange) {
|
||||
rr.setObject(idx, Float.NaN);
|
||||
idx += 1;
|
||||
}
|
||||
} else {
|
||||
DataType dtype = this.columns.get(j).getDataType();
|
||||
String format = this.columns.get(j).getFormat();
|
||||
rr = Array.factory(dtype, new int[]{rn});
|
||||
Array mr = ((List<Array>) this.data).get(j);
|
||||
Object v = DataTypeUtil.convertTo(null, dtype, format);
|
||||
for (int i : rowRange) {
|
||||
if (i < 0) {
|
||||
rr.setObject(idx, v);
|
||||
} else {
|
||||
rr.setObject(idx, mr.getObject(i));
|
||||
}
|
||||
idx += 1;
|
||||
}
|
||||
}
|
||||
((ArrayList) r).add(rr);
|
||||
}
|
||||
if (cols.size() == 1) {
|
||||
r = ((ArrayList) r).get(0);
|
||||
}
|
||||
}
|
||||
|
||||
if (r == null) {
|
||||
return null;
|
||||
} else {
|
||||
Index rIndex = Index.factory(rowKeys);
|
||||
if (cols.size() == 1 && this.columns.size() > 1) {
|
||||
Series s = new Series((Array) r, rIndex, cols.get(0).getName());
|
||||
return s;
|
||||
} else {
|
||||
DataFrame df;
|
||||
if (r instanceof Array) {
|
||||
df = new DataFrame((Array) r, rIndex, cols);
|
||||
} else {
|
||||
df = new DataFrame((ArrayList) r, rIndex, cols);
|
||||
}
|
||||
return df;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Extract DataFrame by row and column ranges
|
||||
*
|
||||
|
||||
@ -34,5 +34,5 @@
|
||||
</Font>
|
||||
<LookFeel DockWindowDecorated="true" LafDecorated="true" Name="FlatDarkLaf"/>
|
||||
<Figure DoubleBuffering="true"/>
|
||||
<Startup MainFormLocation="-7,0" MainFormSize="1408,779"/>
|
||||
<Startup MainFormLocation="-7,0" MainFormSize="1393,791"/>
|
||||
</MeteoInfo>
|
||||
|
||||
Binary file not shown.
@ -801,6 +801,88 @@ class DataFrame(object):
|
||||
ascending = [ascending] * len(by)
|
||||
df = self._dataframe.sortBy(by, ascending)
|
||||
return DataFrame(dataframe=df)
|
||||
|
||||
def reindex(self, index=None, columns=None, axis=None):
|
||||
"""
|
||||
Conform DataFrame to new index with optional filling logic.
|
||||
|
||||
:param index: (*array-like*) New labels for the index. Preferably an Index object to avoid
|
||||
duplicating data.
|
||||
:param columns: (*array-like*) New labels for the columns. Preferably an Index object to
|
||||
avoid duplicating data.
|
||||
:param axis: (*int or str*) Axis to target. Can be either the axis name (‘index’, ‘columns’)
|
||||
or number (0, 1).
|
||||
|
||||
:return: DataFrame with changed index.
|
||||
"""
|
||||
if index is None:
|
||||
index = slice(None)
|
||||
|
||||
k = index
|
||||
rkeys = index
|
||||
if isinstance(k, slice):
|
||||
sidx = 0 if k.start is None else self._index.index(k.start)
|
||||
if sidx < 0:
|
||||
raise KeyError(key)
|
||||
eidx = self.shape[0] - 1 if k.stop is None else self._index.index(k.stop)
|
||||
if eidx < 0:
|
||||
raise KeyError(key)
|
||||
step = 1 if k.step is None else k.step
|
||||
rowkey = Range(sidx, eidx, step)
|
||||
else:
|
||||
rowkey = self._index.get_indexer(k)
|
||||
if len(rowkey) == 0:
|
||||
raise KeyError(key)
|
||||
|
||||
k = columns
|
||||
if k is None:
|
||||
colkey = range(0, self.shape[1], 1)
|
||||
else:
|
||||
if isinstance(k, slice):
|
||||
sidx = 0 if k.start is None else self.columns.indexOfName(k.start)
|
||||
if sidx < 0:
|
||||
raise KeyError(key)
|
||||
eidx = self.shape[1] - 1 if k.stop is None else self.columns.indexOfName(k.stop)
|
||||
if eidx < 0:
|
||||
raise KeyError(key)
|
||||
step = 1 if k.step is None else k.step
|
||||
colkey = Range(sidx, eidx, step)
|
||||
elif isinstance(k, list):
|
||||
colkey = self.columns.indexOfName(k)
|
||||
elif isinstance(k, basestring):
|
||||
col = self.columns.indexOfName(k)
|
||||
if col < 0:
|
||||
raise KeyError(key)
|
||||
colkey = [col]
|
||||
else:
|
||||
return None
|
||||
|
||||
if isinstance(rowkey, (int, Range)):
|
||||
r = self._dataframe.select(rowkey, colkey)
|
||||
else:
|
||||
if isinstance(colkey, Range):
|
||||
ncol = colkey.length()
|
||||
else:
|
||||
ncol = len(colkey)
|
||||
|
||||
if rkeys is None:
|
||||
r = self._dataframe.select(rowkey, colkey)
|
||||
else:
|
||||
if not isinstance(rkeys, list):
|
||||
rkeys = [rkeys]
|
||||
if columns is None:
|
||||
columns = self.columns.names
|
||||
r = self._dataframe.reIndex(rkeys, rowkey, colkey, columns)
|
||||
|
||||
if r is None:
|
||||
return None
|
||||
|
||||
if isinstance(r, MISeries):
|
||||
r = series.Series(series=r)
|
||||
else:
|
||||
r = DataFrame(dataframe=r)
|
||||
|
||||
return r
|
||||
|
||||
def groupby(self, by):
|
||||
"""
|
||||
|
||||
Binary file not shown.
@ -129,6 +129,22 @@ class Index(object):
|
||||
def fill_keylist(self, rdata, rfdata):
|
||||
return self._index.fillKeyList(rdata.asarray(), rfdata)
|
||||
|
||||
def get_indexer(self, key):
|
||||
"""
|
||||
Compute indexer and mask for new index given the current index.
|
||||
|
||||
:param key: Index.
|
||||
|
||||
:return: (*array*) Integers from 0 to n - 1 indicating that the index at these positions matches
|
||||
the corresponding target values. Missing values in the target are marked by -1.
|
||||
"""
|
||||
if isinstance(key, np.NDArray):
|
||||
r = self._index.getIndices(key.asarray())
|
||||
else:
|
||||
r = self._index.getIndices(key)
|
||||
|
||||
return list(r[2])
|
||||
|
||||
def get_format(self):
|
||||
"""
|
||||
Get value to string format.
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user