{{ header }}
This is a repository for short and sweet examples and links for useful pandas recipes. We encourage users to add to this documentation.
Adding interesting links and/or inline examples to this section is a great First Pull Request.
Simplified, condensed, new-user friendly, in-line examples have been inserted where possible to augment the Stack-Overflow and GitHub links. Many of the links contain expanded information, above what the in-line examples offer.
pandas (pd) and NumPy (np) are the only two abbreviated imported modules. The rest are kept explicitly imported for newer users.
These are some neat pandas idioms
if-then/if-then-else on one column, and assignment to another one or more columns:
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
An if-then on one column
.. ipython:: python df.loc[df.AAA >= 5, "BBB"] = -1 df
An if-then with assignment to 2 columns:
.. ipython:: python df.loc[df.AAA >= 5, ["BBB", "CCC"]] = 555 df
Add another line with different logic, to do the -else
.. ipython:: python df.loc[df.AAA < 5, ["BBB", "CCC"]] = 2000 df
Or use pandas where after you've set up a mask
.. ipython:: python
df_mask = pd.DataFrame(
{"AAA": [True] * 4, "BBB": [False] * 4, "CCC": [True, False] * 2}
)
df.where(df_mask, -1000)
if-then-else using NumPy's where()
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
df["logic"] = np.where(df["AAA"] > 5, "high", "low")
df
Split a frame with a boolean criterion
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
df[df.AAA <= 5]
df[df.AAA > 5]
Select with multi-column criteria
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
...and (without assignment returns a Series)
.. ipython:: python df.loc[(df["BBB"] < 25) & (df["CCC"] >= -40), "AAA"]
...or (without assignment returns a Series)
.. ipython:: python df.loc[(df["BBB"] > 25) | (df["CCC"] >= -40), "AAA"]
...or (with assignment modifies the DataFrame.)
.. ipython:: python df.loc[(df["BBB"] > 25) | (df["CCC"] >= 75), "AAA"] = 999 df
Select rows with data closest to certain value using argsort
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
aValue = 43.0
df.loc[(df.CCC - aValue).abs().argsort()]
Dynamically reduce a list of criteria using a binary operators
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
Crit1 = df.AAA <= 5.5
Crit2 = df.BBB == 10.0
Crit3 = df.CCC > -40.0
One could hard code:
.. ipython:: python AllCrit = Crit1 & Crit2 & Crit3
...Or it can be done with a list of dynamically built criteria
.. ipython:: python import functools CritList = [Crit1, Crit2, Crit3] AllCrit = functools.reduce(lambda x, y: x & y, CritList) df[AllCrit]
The :ref:`indexing <indexing>` docs.
Using both row labels and value conditionals
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
df[(df.AAA <= 6) & (df.index.isin([0, 2, 4]))]
Use loc for label-oriented slicing and iloc positional slicing :issue:`2904`
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]},
index=["foo", "bar", "boo", "kar"],
)
There are 2 explicit slicing methods, with a third general case
- Positional-oriented (Python slicing style : exclusive of end)
- Label-oriented (Non-Python slicing style : inclusive of end)
- General (Either slicing style : depends on if the slice contains labels or positions)
.. ipython:: python df.iloc[0:3] # Positional df.loc["bar":"kar"] # Label # Generic df[0:3] df["bar":"kar"]
Ambiguity arises when an index consists of integers with a non-zero start or non-unit increment.
.. ipython:: python
data = {"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
df2 = pd.DataFrame(data=data, index=[1, 2, 3, 4]) # Note index starts at 1.
df2.iloc[1:3] # Position-oriented
df2.loc[1:3] # Label-oriented
Using inverse operator (~) to take the complement of a mask
.. ipython:: python
df = pd.DataFrame(
{"AAA": [4, 5, 6, 7], "BBB": [10, 20, 30, 40], "CCC": [100, 50, -30, -50]}
)
df
df[~((df.AAA <= 6) & (df.index.isin([0, 2, 4])))]
Efficiently and dynamically creating new columns using DataFrame.map (previously named applymap)
.. ipython:: python
df = pd.DataFrame({"AAA": [1, 2, 1, 3], "BBB": [1, 1, 2, 2], "CCC": [2, 1, 3, 1]})
df
source_cols = df.columns # Or some subset would work too
new_cols = [str(x) + "_cat" for x in source_cols]
categories = {1: "Alpha", 2: "Beta", 3: "Charlie"}
df[new_cols] = df[source_cols].map(categories.get)
df
Keep other columns when using min() with groupby
.. ipython:: python
df = pd.DataFrame(
{"AAA": [1, 1, 1, 2, 2, 2, 3, 3], "BBB": [2, 1, 3, 4, 5, 1, 2, 3]}
)
df
Method 1 : idxmin() to get the index of the minimums
.. ipython:: python
df.loc[df.groupby("AAA")["BBB"].idxmin()]
Method 2 : sort then take first of each
.. ipython:: python
df.sort_values(by="BBB").groupby("AAA", as_index=False).first()
Notice the same results, with the exception of the index.
The :ref:`multindexing <advanced.hierarchical>` docs.
Creating a MultiIndex from a labeled frame
.. ipython:: python
df = pd.DataFrame(
{
"row": [0, 1, 2],
"One_X": [1.1, 1.1, 1.1],
"One_Y": [1.2, 1.2, 1.2],
"Two_X": [1.11, 1.11, 1.11],
"Two_Y": [1.22, 1.22, 1.22],
}
)
df
# As Labelled Index
df = df.set_index("row")
df
# With Hierarchical Columns
df.columns = pd.MultiIndex.from_tuples([tuple(c.split("_")) for c in df.columns])
df
# Now stack & Reset
df = df.stack(0, future_stack=True).reset_index(1)
df
# And fix the labels (Notice the label 'level_1' got added automatically)
df.columns = ["Sample", "All_X", "All_Y"]
df
Performing arithmetic with a MultiIndex that needs broadcasting
.. ipython:: python
cols = pd.MultiIndex.from_tuples(
[(x, y) for x in ["A", "B", "C"] for y in ["O", "I"]]
)
df = pd.DataFrame(np.random.randn(2, 6), index=["n", "m"], columns=cols)
df
df = df.div(df["C"], level=1)
df
.. ipython:: python
coords = [("AA", "one"), ("AA", "six"), ("BB", "one"), ("BB", "two"), ("BB", "six")]
index = pd.MultiIndex.from_tuples(coords)
df = pd.DataFrame([11, 22, 33, 44, 55], index, ["MyData"])
df
To take the cross section of the 1st level and 1st axis the index:
.. ipython:: python
# Note : level and axis are optional, and default to zero
df.xs("BB", level=0, axis=0)
...and now the 2nd level of the 1st axis.
.. ipython:: python
df.xs("six", level=1, axis=0)
Slicing a MultiIndex with xs, method #2
.. ipython:: python
import itertools
index = list(itertools.product(["Ada", "Quinn", "Violet"], ["Comp", "Math", "Sci"]))
headr = list(itertools.product(["Exams", "Labs"], ["I", "II"]))
indx = pd.MultiIndex.from_tuples(index, names=["Student", "Course"])
cols = pd.MultiIndex.from_tuples(headr) # Notice these are un-named
data = [[70 + x + y + (x * y) % 3 for x in range(4)] for y in range(9)]
df = pd.DataFrame(data, indx, cols)
df
All = slice(None)
df.loc["Violet"]
df.loc[(All, "Math"), All]
df.loc[(slice("Ada", "Quinn"), "Math"), All]
df.loc[(All, "Math"), ("Exams")]
df.loc[(All, "Math"), (All, "II")]
Setting portions of a MultiIndex with xs
Sort by specific column or an ordered list of columns, with a MultiIndex
.. ipython:: python
df.sort_values(by=("Labs", "II"), ascending=False)
Partial selection, the need for sortedness :issue:`2995`
Prepending a level to a multiindex
The :ref:`missing data<missing_data>` docs.
Fill forward a reversed timeseries
.. ipython:: python
df = pd.DataFrame(
np.random.randn(6, 1),
index=pd.date_range("2013-08-01", periods=6, freq="B"),
columns=list("A"),
)
df.loc[df.index[3], "A"] = np.nan
df
df.bfill()
The :ref:`grouping <groupby>` docs.
Unlike agg, apply's callable is passed a sub-DataFrame which gives you access to all the columns
.. ipython:: python
df = pd.DataFrame(
{
"animal": "cat dog cat fish dog cat cat".split(),
"size": list("SSMMMLL"),
"weight": [8, 10, 11, 1, 20, 12, 12],
"adult": [False] * 5 + [True] * 2,
}
)
df
# List the size of the animals with the highest weight.
df.groupby("animal").apply(lambda subf: subf["size"][subf["weight"].idxmax()], include_groups=False)
.. ipython:: python
gb = df.groupby("animal")
gb.get_group("cat")
Apply to different items in a group
.. ipython:: python
def GrowUp(x):
avg_weight = sum(x[x["size"] == "S"].weight * 1.5)
avg_weight += sum(x[x["size"] == "M"].weight * 1.25)
avg_weight += sum(x[x["size"] == "L"].weight)
avg_weight /= len(x)
return pd.Series(["L", avg_weight, True], index=["size", "weight", "adult"])
expected_df = gb.apply(GrowUp, include_groups=False)
expected_df
.. ipython:: python
S = pd.Series([i / 100.0 for i in range(1, 11)])
def cum_ret(x, y):
return x * (1 + y)
def red(x):
return functools.reduce(cum_ret, x, 1.0)
S.expanding().apply(red, raw=True)
Replacing some values with mean of the rest of a group
.. ipython:: python
df = pd.DataFrame({"A": [1, 1, 2, 2], "B": [1, -1, 1, 2]})
gb = df.groupby("A")
def replace(g):
mask = g < 0
return g.where(~mask, g[~mask].mean())
gb.transform(replace)
Sort groups by aggregated data
.. ipython:: python
df = pd.DataFrame(
{
"code": ["foo", "bar", "baz"] * 2,
"data": [0.16, -0.21, 0.33, 0.45, -0.59, 0.62],
"flag": [False, True] * 3,
}
)
code_groups = df.groupby("code")
agg_n_sort_order = code_groups[["data"]].transform("sum").sort_values(by="data")
sorted_df = df.loc[agg_n_sort_order.index]
sorted_df
Create multiple aggregated columns
.. ipython:: python
rng = pd.date_range(start="2014-10-07", periods=10, freq="2min")
ts = pd.Series(data=list(range(10)), index=rng)
def MyCust(x):
if len(x) > 2:
return x.iloc[1] * 1.234
return pd.NaT
mhc = {"Mean": "mean", "Max": "max", "Custom": MyCust}
ts.resample("5min").apply(mhc)
ts
Create a value counts column and reassign back to the DataFrame
.. ipython:: python
df = pd.DataFrame(
{"Color": "Red Red Red Blue".split(), "Value": [100, 150, 50, 50]}
)
df
df["Counts"] = df.groupby(["Color"]).transform(len)
df
Shift groups of the values in a column based on the index
.. ipython:: python
df = pd.DataFrame(
{"line_race": [10, 10, 8, 10, 10, 8], "beyer": [99, 102, 103, 103, 88, 100]},
index=[
"Last Gunfighter",
"Last Gunfighter",
"Last Gunfighter",
"Paynter",
"Paynter",
"Paynter",
],
)
df
df["beyer_shifted"] = df.groupby(level=0)["beyer"].shift(1)
df
Select row with maximum value from each group
.. ipython:: python
df = pd.DataFrame(
{
"host": ["other", "other", "that", "this", "this"],
"service": ["mail", "web", "mail", "mail", "web"],
"no": [1, 2, 1, 2, 1],
}
).set_index(["host", "service"])
mask = df.groupby(level=0).agg("idxmax")
df_count = df.loc[mask["no"]].reset_index()
df_count
Grouping like Python's itertools.groupby
.. ipython:: python df = pd.DataFrame([0, 1, 0, 1, 1, 1, 0, 1, 1], columns=["A"]) df["A"].groupby((df["A"] != df["A"].shift()).cumsum()).groups df["A"].groupby((df["A"] != df["A"].shift()).cumsum()).cumsum()
Rolling Computation window based on values instead of counts
Create a list of dataframes, split using a delineation based on logic included in rows.
.. ipython:: python
df = pd.DataFrame(
data={
"Case": ["A", "A", "A", "B", "A", "A", "B", "A", "A"],
"Data": np.random.randn(9),
}
)
dfs = list(
zip(
*df.groupby(
(1 * (df["Case"] == "B"))
.cumsum()
.rolling(window=3, min_periods=1)
.median()
)
)
)[-1]
dfs[0]
dfs[1]
dfs[2]
The :ref:`Pivot <reshaping.pivot>` docs.
.. ipython:: python
df = pd.DataFrame(
data={
"Province": ["ON", "QC", "BC", "AL", "AL", "MN", "ON"],
"City": [
"Toronto",
"Montreal",
"Vancouver",
"Calgary",
"Edmonton",
"Winnipeg",
"Windsor",
],
"Sales": [13, 6, 16, 8, 4, 3, 1],
}
)
table = pd.pivot_table(
df,
values=["Sales"],
index=["Province"],
columns=["City"],
aggfunc="sum",
margins=True,
)
table.stack("City", future_stack=True)
Frequency table like plyr in R
.. ipython:: python
grades = [48, 99, 75, 80, 42, 80, 72, 68, 36, 78]
df = pd.DataFrame(
{
"ID": ["x%d" % r for r in range(10)],
"Gender": ["F", "M", "F", "M", "F", "M", "F", "M", "M", "M"],
"ExamYear": [
"2007",
"2007",
"2007",
"2008",
"2008",
"2008",
"2008",
"2009",
"2009",
"2009",
],
"Class": [
"algebra",
"stats",
"bio",
"algebra",
"algebra",
"stats",
"stats",
"algebra",
"bio",
"bio",
],
"Participated": [
"yes",
"yes",
"yes",
"yes",
"no",
"yes",
"yes",
"yes",
"yes",
"yes",
],
"Passed": ["yes" if x > 50 else "no" for x in grades],
"Employed": [
True,
True,
True,
False,
False,
False,
False,
True,
True,
False,
],
"Grade": grades,
}
)
df.groupby("ExamYear").agg(
{
"Participated": lambda x: x.value_counts()["yes"],
"Passed": lambda x: sum(x == "yes"),
"Employed": lambda x: sum(x),
"Grade": lambda x: sum(x) / len(x),
}
)
Plot pandas DataFrame with year over year data
To create year and month cross tabulation:
.. ipython:: python
df = pd.DataFrame(
{"value": np.random.randn(36)},
index=pd.date_range("2011-01-01", freq="M", periods=36),
)
pd.pivot_table(
df, index=df.index.month, columns=df.index.year, values="value", aggfunc="sum"
)
Rolling apply to organize - Turning embedded lists into a MultiIndex frame
.. ipython:: python
df = pd.DataFrame(
data={
"A": [[2, 4, 8, 16], [100, 200], [10, 20, 30]],
"B": [["a", "b", "c"], ["jj", "kk"], ["ccc"]],
},
index=["I", "II", "III"],
)
def make_df(ser):
new_vals = [pd.Series(value, name=name) for name, value in ser.items()]
return pd.DataFrame(new_vals)
df_orgz = pd.concat({ind: row.pipe(make_df) for ind, row in df.iterrows()})
df_orgz
Rolling apply with a DataFrame returning a Series
Rolling Apply to multiple columns where function calculates a Series before a Scalar from the Series is returned
.. ipython:: python
df = pd.DataFrame(
data=np.random.randn(2000, 2) / 10000,
index=pd.date_range("2001-01-01", periods=2000),
columns=["A", "B"],
)
df
def gm(df, const):
v = ((((df["A"] + df["B"]) + 1).cumprod()) - 1) * const
return v.iloc[-1]
s = pd.Series(
{
df.index[i]: gm(df.iloc[i: min(i + 51, len(df) - 1)], 5)
for i in range(len(df) - 50)
}
)
s
Rolling apply with a DataFrame returning a Scalar
Rolling Apply to multiple columns where function returns a Scalar (Volume Weighted Average Price)
.. ipython:: python
rng = pd.date_range(start="2014-01-01", periods=100)
df = pd.DataFrame(
{
"Open": np.random.randn(len(rng)),
"Close": np.random.randn(len(rng)),
"Volume": np.random.randint(100, 2000, len(rng)),
},
index=rng,
)
df
def vwap(bars):
return (bars.Close * bars.Volume).sum() / bars.Volume.sum()
window = 5
s = pd.concat(
[
(pd.Series(vwap(df.iloc[i: i + window]), index=[df.index[i + window]]))
for i in range(len(df) - window)
]
)
s.round(2)
Constructing a datetime range that excludes weekends and includes only certain times
Aggregation and plotting time series
Turn a matrix with hours in columns and days in rows into a continuous row sequence in the form of a time series. How to rearrange a Python pandas DataFrame?
Dealing with duplicates when reindexing a timeseries to a specified frequency
Calculate the first day of the month for each entry in a DatetimeIndex
.. ipython:: python
dates = pd.date_range("2000-01-01", periods=5)
dates.to_period(freq="M").to_timestamp()
The :ref:`Resample <timeseries.resampling>` docs.
Using Grouper instead of TimeGrouper for time grouping of values
Time grouping with some missing values
Valid frequency arguments to Grouper :ref:`Timeseries <timeseries.offset_aliases>`
Using TimeGrouper and another grouping to create subgroups, then apply a custom function :issue:`3791`
Resampling with custom periods
Resample intraday frame without adding new days
The :ref:`Join <merging.join>` docs.
Concatenate two dataframes with overlapping index (emulate R rbind)
.. ipython:: python
rng = pd.date_range("2000-01-01", periods=6)
df1 = pd.DataFrame(np.random.randn(6, 3), index=rng, columns=["A", "B", "C"])
df2 = df1.copy()
Depending on df construction, ignore_index may be needed
.. ipython:: python df = pd.concat([df1, df2], ignore_index=True) df
Self Join of a DataFrame :issue:`2996`
.. ipython:: python
df = pd.DataFrame(
data={
"Area": ["A"] * 5 + ["C"] * 2,
"Bins": [110] * 2 + [160] * 3 + [40] * 2,
"Test_0": [0, 1, 0, 1, 2, 0, 1],
"Data": np.random.randn(7),
}
)
df
df["Test_1"] = df["Test_0"] - 1
pd.merge(
df,
df,
left_on=["Bins", "Area", "Test_0"],
right_on=["Bins", "Area", "Test_1"],
suffixes=("_L", "_R"),
)
Join with a criteria based on the values
Using searchsorted to merge based on values inside a range
The :ref:`Plotting <visualization>` docs.
Setting x-axis major and minor labels
Plotting multiple charts in an IPython Jupyter notebook
Annotate a time-series plot #2
Generate Embedded plots in excel files using Pandas, Vincent and xlsxwriter
Boxplot for each quartile of a stratifying variable
.. ipython:: python
df = pd.DataFrame(
{
"stratifying_var": np.random.uniform(0, 100, 20),
"price": np.random.normal(100, 5, 20),
}
)
df["quartiles"] = pd.qcut(
df["stratifying_var"], 4, labels=["0-25%", "25-50%", "50-75%", "75-100%"]
)
@savefig quartile_boxplot.png
df.boxplot(column="price", by="quartiles")
Performance comparison of SQL vs HDF5
The :ref:`CSV <io.read_csv_table>` docs
Reading only certain rows of a csv chunk-by-chunk
Reading the first few lines of a frame
Reading a file that is compressed but not by gzip/bz2 (the native compressed formats which read_csv understands).
This example shows a WinZipped file, but is a general application of opening the file within a context manager and
using that handle to read.
See here
Dealing with bad lines :issue:`2886`
Write a multi-row index CSV without writing duplicates
The best way to combine multiple files into a single DataFrame is to read the individual frames one by one, put all of the individual frames into a list, and then combine the frames in the list using :func:`pd.concat`:
.. ipython:: python
for i in range(3):
data = pd.DataFrame(np.random.randn(10, 4))
data.to_csv("file_{}.csv".format(i))
files = ["file_0.csv", "file_1.csv", "file_2.csv"]
result = pd.concat([pd.read_csv(f) for f in files], ignore_index=True)
You can use the same approach to read all files matching a pattern. Here is an example using glob:
.. ipython:: python
import glob
import os
files = glob.glob("file_*.csv")
result = pd.concat([pd.read_csv(f) for f in files], ignore_index=True)
Finally, this strategy will work with the other pd.read_*(...) functions described in the :ref:`io docs<io>`.
.. ipython:: python
:suppress:
for i in range(3):
os.remove("file_{}.csv".format(i))
Parsing date components in multi-columns is faster with a format
.. ipython:: python
i = pd.date_range("20000101", periods=10000)
df = pd.DataFrame({"year": i.year, "month": i.month, "day": i.day})
df.head()
%timeit pd.to_datetime(df.year * 10000 + df.month * 100 + df.day, format='%Y%m%d')
ds = df.apply(lambda x: "%04d%02d%02d" % (x["year"], x["month"], x["day"]), axis=1)
ds.head()
%timeit pd.to_datetime(ds)
.. ipython:: python
data = """;;;;
;;;;
;;;;
;;;;
;;;;
;;;;
;;;;
;;;;
;;;;
;;;;
date;Param1;Param2;Param4;Param5
;m²;°C;m²;m
;;;;
01.01.1990 00:00;1;1;2;3
01.01.1990 01:00;5;3;4;5
01.01.1990 02:00;9;5;6;7
01.01.1990 03:00;13;7;8;9
01.01.1990 04:00;17;9;10;11
01.01.1990 05:00;21;11;12;13
"""
.. ipython:: python
from io import StringIO
pd.read_csv(
StringIO(data),
sep=";",
skiprows=[11, 12],
index_col=0,
parse_dates=True,
header=10,
)
.. ipython:: python
pd.read_csv(StringIO(data), sep=";", header=10, nrows=10).columns
columns = pd.read_csv(StringIO(data), sep=";", header=10, nrows=10).columns
pd.read_csv(
StringIO(data), sep=";", index_col=0, header=12, parse_dates=True, names=columns
)
The :ref:`SQL <io.sql>` docs
Reading from databases with SQL
The :ref:`Excel <io.excel>` docs
Reading from a filelike handle
Modifying formatting in XlsxWriter output
Loading only visible sheets :issue:`19842#issuecomment-892150745`
Reading HTML tables from a server that cannot handle the default request header
The :ref:`HDFStores <io.hdf5>` docs
Simple queries with a Timestamp Index
Managing heterogeneous data using a linked multiple table hierarchy :issue:`3032`
Merging on-disk tables with millions of rows
Avoiding inconsistencies when writing to a store from multiple processes/threads
De-duplicating a large store by chunks, essentially a recursive reduction operation. Shows a function for taking in data from csv file and creating a store by chunks, with date parsing as well. See here
Creating a store chunk-by-chunk from a csv file
Appending to a store, while creating a unique index
Reading in a sequence of files, then providing a global unique index to a store while appending
Groupby on a HDFStore with low group density
Groupby on a HDFStore with high group density
Hierarchical queries on a HDFStore
Troubleshoot HDFStore exceptions
Setting min_itemsize with strings
Using ptrepack to create a completely-sorted-index on a store
Storing Attributes to a group node
.. ipython:: python
df = pd.DataFrame(np.random.randn(8, 3))
store = pd.HDFStore("test.h5")
store.put("df", df)
# you can store an arbitrary Python object via pickle
store.get_storer("df").attrs.my_attribute = {"A": 10}
store.get_storer("df").attrs.my_attribute
.. ipython:: python
:suppress:
store.close()
os.remove("test.h5")
You can create or load a HDFStore in-memory by passing the driver
parameter to PyTables. Changes are only written to disk when the HDFStore
is closed.
.. ipython:: python
store = pd.HDFStore("test.h5", "w", driver="H5FD_CORE")
df = pd.DataFrame(np.random.randn(8, 3))
store["test"] = df
# only after closing the store, data is written to disk:
store.close()
.. ipython:: python
:suppress:
os.remove("test.h5")
pandas readily accepts NumPy record arrays, if you need to read in a binary
file consisting of an array of C structs. For example, given this C program
in a file called main.c compiled with gcc main.c -std=gnu99 on a
64-bit machine,
#include <stdio.h>
#include <stdint.h>
typedef struct _Data
{
int32_t count;
double avg;
float scale;
} Data;
int main(int argc, const char *argv[])
{
size_t n = 10;
Data d[n];
for (int i = 0; i < n; ++i)
{
d[i].count = i;
d[i].avg = i + 1.0;
d[i].scale = (float) i + 2.0f;
}
FILE *file = fopen("binary.dat", "wb");
fwrite(&d, sizeof(Data), n, file);
fclose(file);
return 0;
}the following Python code will read the binary file 'binary.dat' into a
pandas DataFrame, where each element of the struct corresponds to a column
in the frame:
names = "count", "avg", "scale"
# note that the offsets are larger than the size of the type because of
# struct padding
offsets = 0, 8, 16
formats = "i4", "f8", "f4"
dt = np.dtype({"names": names, "offsets": offsets, "formats": formats}, align=True)
df = pd.DataFrame(np.fromfile("binary.dat", dt))Note
The offsets of the structure elements may be different depending on the architecture of the machine on which the file was created. Using a raw binary file format like this for general data storage is not recommended, as it is not cross platform. We recommended either HDF5 or parquet, both of which are supported by pandas' IO facilities.
Numerical integration (sample-based) of a time series
Often it's useful to obtain the lower (or upper) triangular form of a correlation matrix calculated from :func:`DataFrame.corr`. This can be achieved by passing a boolean mask to where as follows:
.. ipython:: python
df = pd.DataFrame(np.random.random(size=(100, 5)))
corr_mat = df.corr()
mask = np.tril(np.ones_like(corr_mat, dtype=np.bool_), k=-1)
corr_mat.where(mask)
The method argument within DataFrame.corr can accept a callable in addition to the named correlation types. Here we compute the distance correlation matrix for a DataFrame object.
.. ipython:: python
def distcorr(x, y):
n = len(x)
a = np.zeros(shape=(n, n))
b = np.zeros(shape=(n, n))
for i in range(n):
for j in range(i + 1, n):
a[i, j] = abs(x[i] - x[j])
b[i, j] = abs(y[i] - y[j])
a += a.T
b += b.T
a_bar = np.vstack([np.nanmean(a, axis=0)] * n)
b_bar = np.vstack([np.nanmean(b, axis=0)] * n)
A = a - a_bar - a_bar.T + np.full(shape=(n, n), fill_value=a_bar.mean())
B = b - b_bar - b_bar.T + np.full(shape=(n, n), fill_value=b_bar.mean())
cov_ab = np.sqrt(np.nansum(A * B)) / n
std_a = np.sqrt(np.sqrt(np.nansum(A ** 2)) / n)
std_b = np.sqrt(np.sqrt(np.nansum(B ** 2)) / n)
return cov_ab / std_a / std_b
df = pd.DataFrame(np.random.normal(size=(100, 3)))
df.corr(method=distcorr)
The :ref:`Timedeltas <timedeltas.timedeltas>` docs.
.. ipython:: python
import datetime
s = pd.Series(pd.date_range("2012-1-1", periods=3, freq="D"))
s - s.max()
s.max() - s
s - datetime.datetime(2011, 1, 1, 3, 5)
s + datetime.timedelta(minutes=5)
datetime.datetime(2011, 1, 1, 3, 5) - s
datetime.timedelta(minutes=5) + s
Adding and subtracting deltas and dates
.. ipython:: python
deltas = pd.Series([datetime.timedelta(days=i) for i in range(3)])
df = pd.DataFrame({"A": s, "B": deltas})
df
df["New Dates"] = df["A"] + df["B"]
df["Delta"] = df["A"] - df["New Dates"]
df
df.dtypes
Values can be set to NaT using np.nan, similar to datetime
.. ipython:: python y = s - s.shift() y y[1] = np.nan y
To create a dataframe from every combination of some given values, like R's expand.grid()
function, we can create a dict where the keys are column names and the values are lists
of the data values:
.. ipython:: python
def expand_grid(data_dict):
rows = itertools.product(*data_dict.values())
return pd.DataFrame.from_records(rows, columns=data_dict.keys())
df = expand_grid(
{"height": [60, 70], "weight": [100, 140, 180], "sex": ["Male", "Female"]}
)
df
To assess if a series has a constant value, we can check if series.nunique() <= 1.
However, a more performant approach, that does not count all unique values first, is:
.. ipython:: python v = s.to_numpy() is_constant = v.shape[0] == 0 or (s[0] == s).all()
This approach assumes that the series does not contain missing values. For the case that we would drop NA values, we can simply remove those values first:
.. ipython:: python v = s.dropna().to_numpy() is_constant = v.shape[0] == 0 or (s[0] == s).all()
If missing values are considered distinct from any other value, then one could use:
.. ipython:: python v = s.to_numpy() is_constant = v.shape[0] == 0 or (s[0] == s).all() or not pd.notna(v).any()
(Note that this example does not disambiguate between np.nan, pd.NA and None)