I've been working on this as a beginner for a while. Overall, I want to read in a NetCDF file and import multiple (~50) columns (and 17520 cases) into a Pandas DataFrame. At the moment I have set it up for a list of 4 variables but I want to be able to expand that somehow. I made a start, but any help on how to loop through to make this happen with 50 variables would be great. It does work using the code below for 4 variables. I know its not pretty - still learning!
Another question I have it that when I try to read the numpy arrays directly into Pandas DataFrame it doesn't work and instead creates a DataFrame that is 17520 columns large. It should be the other way (transposed). If I create a series, it works fine. So I have had to use the following lines to get around this. Not even sure why it works. Any suggestions of a better way (especially when it comes to 50 variables)?
d={vnames[0] :vartemp[0], vnames[1] :vartemp[1], vnames[2] :vartemp[2], vnames[3] :vartemp[3]}
hs = pd.DataFrame(d,index=times)
The whole code is pasted below:
import pandas as pd
import datetime as dt
import xlrd
import numpy as np
import netCDF4
def excel_to_pydate(exceldate):
datemode=0 # datemode: 0 for 1900-based, 1 for 1904-based
pyear, pmonth, pday, phour, pminute, psecond = xlrd.xldate_as_tuple(exceldate, datemode)
py_date = dt.datetime(pyear, pmonth, pday, phour, pminute, psecond)
return(py_date)
def main():
filename='HowardSprings_2010_L4.nc'
#Define a list of variables names we want from the netcdf file
vnames = ['xlDateTime', 'Fa', 'Fh' ,'Fg']
# Open the NetCDF file
nc = netCDF4.Dataset(filename)
#Create some lists of size equal to length of vnames list.
temp=list(xrange(len(vnames)))
vartemp=list(xrange(len(vnames)))
#Enumerate the list and assign each NetCDF variable to an element in the lists.
# First get the netcdf variable object assign to temp
# Then strip the data from that and add to temporary variable (vartemp)
for index, variable in enumerate(vnames):
temp[index]= nc.variables[variable]
vartemp[index] = temp[index][:]
# Now call the function to convert to datetime from excel. Assume datemode: 0
times = [excel_to_pydate(elem) for elem in vartemp[0]]
#Dont know why I cant just pass a list of variables i.e. [vartemp[0], vartemp[1], vartemp[2]]
#But this is only thing that worked
#Create Pandas dataframe using times as index
d={vnames[0] :vartemp[0], vnames[1] :vartemp[1], vnames[2] :vartemp[2], vnames[3] :vartemp[3]}
theDataFrame = pd.DataFrame(d,index=times)
#Define missing data value and apply to DataFrame
missing=-9999
theDataFrame1=theDataFrame.replace({vnames[0] :missing, vnames[1] :missing, vnames[2] :missing, vnames[3] :missing},'NaN')
main()
You could replace:
d = {vnames[0] :vartemp[0], ..., vnames[3]: vartemp[3]}
hs = pd.DataFrame(d, index=times)
with
hs = pd.DataFrame(vartemp[0:4], columns=vnames[0:4], index=times)
.
Saying that, pandas can read HDF5 directly, so perhaps the same is true for netCDF (which is based on HDF5)...
Related
I have many CSV files saved in AWS s3 with the same first set of columns and a lot of optional columns. I don't want to download them one by one and than use pd.concat to read it, since this takes a lot of time and has to fit in to the computer memory. Instead, I'm trying to use Dask to load and sum up all of these files, when optional columns should should be treated as zeros.
If all columns where the same I could use:
import dask.dataframe as dd
addr = "s3://SOME_BASE_ADDRESS/*.csv"
df = dd.read_csv(addr)
df.groupby(["index"]).sum().compute()
but it doesn't work with files that don't have same number of columns, since Dask assumes it can use the first columns for all files:
File ".../lib/python3.7/site-packages/pandas/core/internals/managers.py", line 155, in set_axis
'values have {new} elements'.format(old=old_len, new=new_len))
ValueError: Length mismatch: Expected axis has 64 elements, new values have 62 elements
According to this thread I can either read all headers in advanced (for example by writing them as I produce and save all of the small CSV's) or use something like this:
df = dd.concat([dd.read_csv(f) for f in filelist])
I wonder if this solution is actually faster/better than just directly use pandas? In general, I'd like to know what is the best (mainly fastest) way to tackle this issue?
It might be a good idea to use delayed to standardize dataframes before converting them to a dask dataframe (whether this is optimal for your use case is difficult to judge).
import dask.dataframe as dd
from dask import delayed
list_files = [...] # create a list of files inside s3 bucket
list_cols_to_keep = ['col1', 'col2']
#delayed
def standard_csv(file_path):
df = pd.read_csv(file_path)
df = df[list_cols_to_keep]
# add any other standardization routines, e.g. dtype conversion
return df
ddf = dd.from_delayed([standard_csv(f) for f in list_files])
I ended up giving up using Dask since it was too slow and used aws s3 sync to download the data and multiprocessing.Pool to read and concat them:
# download:
def sync_outputs(out_path):
local_dir_path = f"/tmp/outputs/"
safe_mkdir(os.path.dirname(local_dir_path))
cmd = f'aws s3 sync {job_output_dir} {local_dir_path} > /tmp/null' # the last part is to avoid prints
os.system(cmd)
return local_dir_path
# concat:
def read_csv(path):
return pd.read_csv(path,index_col=0)
def read_csvs_parallel(local_paths):
from multiprocessing import Pool
import os
with Pool(os.cpu_count()) as p:
csvs = list(tqdm(p.imap(read_csv, local_paths), desc='reading csvs', total=len(paths)))
return csvs
# all togeter:
def concat_csvs_parallel(out_path):
local_paths = sync_outputs(out_path)
csvs = read_csvs_parallel(local_paths)
df = pd.concat(csvs)
return df
aws s3 sync dowloaded about 1000 files (~1KB each) in about 30 second, and reading than with multiproccesing (8 cores) took 3 seconds, this was much faster than also downloading the files using multiprocessing (almost 2 minutes for 1000 files)
I have gone days trying to figure this out, hopefully someone can help.
I am uploading a .mat file into python using scipy.io, placing the struct into a dataframe, which will then be used in Tensorflow.
from scipy.io import loadmat
import pandas as pd
import numpy as p
import matplotlib.pyplot as plt
#import TF
path = '/home/anthony/PycharmProjects/Deep_Learning_MATLAB/circuit-data/for tinghao/template1-lib5-eqns-CR-RESULTS-SET1-FINAL.mat'
raw_data = loadmat(path, squeeze_me=True)
data = raw_data['Graphs']
df = pd.DataFrame(data, dtype=int)
df.pop('transferFunc')
print(df.dtypes)
The out put is:
A object
Ln object
types object
nz int64
np int64
dtype: object
Process finished with exit code 0
The struct is (43249x6). Each cell in the 'A' column is a different sized matrix, i.e. 18x18, or 16x16 etc. Each cell in "Ln" is a row of letters each in their own separate cell. Each cell in 'Types' contains 12 columns of numbers, and 'nz' and 'np' i have no issues with.
I want to put all columns into a dataframe, and use column A or LN or Types as the 'Labels' and nz and np as 'features', again i do not have issues with the latter. Can anyone help with this or have some kind of work around.
The end goal is to have tensorflow train on nz and np and give me either a matrix, Ln, or Type.
What type of data is your .mat file of ? Is your application very time critical?
If you can collect all your data in a struct you could give jsonencode a try, make the struct a json file and load it back into python via json (see json documentation on loading data).
Then you can create a pandas dataframe via
pd.df.from_dict()
Of course this would only be a workaround. Still you would have to ensure your data in the MATLAB struct is correctly orderer to be then imported and transferred to a df.
raw_data = loadmat(path, squeeze_me=True)
data = raw_data['Graphs']
graph_labels = pd.DataFrame()
graph_labels['perf'] = raw_data['Objective'][0:1000]
graph_labels['np'] = data['np'][0:1000]
The code above helped out. Its very simple and drawn out, but it got the job done. But, it does not work in tensorflow because tensorflow does not accept this format, and that was my main issue. I have to convert adjacency matrices to networkx graphs, then upload them into stellargraph.
I'm SUPER green to Python and am having some issues trying to automate some calculations.
I know that this works to add a new column called "Returns" that divides "value" of current to "value" of previous to a csv:
import pandas as pd
import numpy as np
import csv
a = pd.read_csv("/Data/a_data.csv", index_col = "time")
a ["Returns"] = (a["value"]/a["value"].shift(1) -1)*100
However, I have a lot of these CSVs. I need this calculation to happen prior to merging them all together. So I was hoping to write something that just looped through all of the CSVs and did the calculation and added the column but clearly this was incorrect as I get Syntax error:
import pandas as pd
import numpy as np
import csv
a = pd.read_csv("/Data/a_data.csv", index_col = "time")
b = pd.read_csv("/Data/b_data.csv", index_col = "time")
c = pd.read_csv("/Data/c_data.csv", index_col = "time")
my_lists = ['a','b','c']
for my_list in my_lists:
{my_list}["Returns"] = ({my_list}["close"]/{my_list}["close"].shift(1) -1)*100
print(f"Calculating: {my_list.upper()}")
I'm sure there is an easy way to do this that I just haven't reached in my Python education yet, so any guidance would be greatly appreciated!
Assuming "close" and "time" are fields defined in each of your csv files, you could define a function that reads each file, do the shift and returns a dataframe:
def your_func(my_file): # this function takes a file name as an argument.
my_df = pd.read_csv(my_file, index_col = "time") # The function reads its content into a data frame,
my_df["Returns"] = (my_df["close"]/{my_df}["close"].shift(1) -1)*100 # makes the calculation
return my_df #and returns it as an output.
Then as a main code, you collect all csv files from a folder with glob package. Using the above function, you build a data frame for each file with the calculation done.
import glob
path =r'/Data/' # path to the directory where you have the csv files
filenames = glob.glob(path + "/*.csv") # grab the csv files names using glob package with path+all csv files present
for filename in filenames: # loop into all csv files names in the list of csv files present in the directory
df= your_func (filename) # call the function, defined above block of code, that reads the file from its name as argument, then makes the calculation and returns it.
print (df)
Above, there is a print of the data Frame which shows results; I am not sure what you intend to do with upper (I dont think this is a function on a data frame).
Finally, this returns independent data frames with calculations done prior to other or final transformation.
1.Do a, b, c data frames have the same dimension?
2.You don't need to import the CSV library because it includes in the Pandas library.
3.If you want to union data frames, you can use like this :
my_lists = [a,b,c]
and you can concatenate with this way:
result=pd.concat(my_lists)
Lastly, your calculation should be :
result["Returns"]=(result.loc[:, "close"].div(result.loc[:, "close"].shift()).fillna(0).replace([np.inf, -np.inf], 0))
You need to add an index-label selection (loc) function to the data frame in order to access the values. When numbers are dividing, results can be NaN(Not a Number) or infinite. Therefore, replace and fillna functions are related to NaN and Inf.
I am importing 4000+ csv files all with the same columns, columns=['Date', 'Datapint'] the importing the csv's to dask is pretty straight forward and is working fine for me.
file_paths = '/root/data/daily/'
df = dd.read_csv(file_paths+'*.csv',
delim_whitespace=True,
names=['Date','Datapoint'])
The task I am trying to achive is to be able to name the 'Datapoint' column the filename of the .csv. I know you can set a column to the path using include_path_column = True. But I am wondering if there is a simple way use that pathname as a column name with out having to run a separate step down the line.
I was able to do this (fairly straight forward) using dask's delayed function:
import pandas as pd
import dask.dataframe as dd
from dask import delayed
import glob
path = r'/root/data/daily' # use your path
file_list = glob.glob(path + "/*.csv")
def read_and_label_csv(filename):
# reads each csv file to a pandas.DataFrame
df_csv = pd.read_csv(filename,
delim_whitespace=True,
names=['Date','Close'])
df_csv.rename(columns={'Close':path_2_column}, inplace=True)
return df_csv
# create a list of functions ready to return a pandas.DataFrame
dfs = [delayed(read_and_label_csv)(fname) for fname in file_list]
# using delayed, assemble the pandas.DataFrames into a dask.DataFrame
ddf = dd.from_delayed(dfs)
It is unclear to me what exactly you are trying to accomplish. If you are just trying to change the name of the column that the filepaths are written to, you can set include_path_column='New Column Name'. If you are naming a column based on the path to each file, it seems like you'll get a rather sparse array once the data are concatenated and I would argue that a groupby would probably work better.
I am trying to read and later on to plot data from a netcdf file. Some of the arrays contained at the .nc file that I am trying to store as variables, are created as a (1,1,n) size variable. When printing them i see [[[ numbers, numbers,....]]]. Why are these three [[[ are created? How can I read these variables as a simple (n,1) array?
Here is my code
import pandas as pd
import netCDF4 as nc
import matplotlib.pyplot as plt
from tkinter import filedialog
import numpy as np
file_path=filedialog.askopenfilename(title = "Select files", filetypes = (("all files","*.*"),("txt files","*.txt")))
file=nc.Dataset(file_path)
print(file.variables.keys()) # get all variable names
read_alt=file.variables['altitude'][:]
alt=np.array(read_alt)
read_b355=file.variables['backscatter'][:]
read_error_b355=file.variables['error_backscatter'][:]
b355=np.array(read_b355)
error_b355=np.array(read_error_b355)
the variable alt is fine, for the other two I have the aforementioned problem.
Is it possible that your variables - altitude, backscatter and error_backscatter - have more than one dimensions? Whenever you load that kind of data, the number of dimensions is kept by the netCDF library.
Nevertheless, what I usually do, is that I remove the dimensions that I do not need from the arrays by squeezing them:
read_alt = np.squeeze(file.variables['altitude'][:])
read_b355 = np.squeeze(file.variables['backscatter'][:]);
read_error_b355 = np.squeeze(file.variables['error_backscatter'][:]);