将 SAV 转换为 CSV

Question

I tried to convert a SAV file into CSV by using the following code

data = pd.io.stata.read_stata("C:/Users/Nicola/Desktop/Relevant projects activities ACF/BRACED Final Evaluation/Evaluations/CSI_compil_2017.sav")
writer = pd.ExcelWriter('C:/Users/Nicola/Desktop/Baseline.xlsx')
data.to_excel(writer, 'data')
data.to_csv('changed_to_csv.csv')
writer.save()

The output I am getting is the following

ValueError: Version of given Stata file is not 104, 105, 108, 111 (Stata 7SE), 113 (Stata 8/9), 114 (Stata 10/11), 115 (Stata 12), 117 (Stata 13), or 118 (Stata 14)

Is there a better snippet I could use to perform this conversion more efficiently? Thanks

Answer 1

Please see the following answer : https://stackoverflow.com/a/20873154/5999386

In short, using import pandas.rpy.common as com to use R capabilities to parse .sav file into Pandas's data frame.

Answer 2

I just managed to convert it with this snippet (R Kernel):

library(foreign)
write.table(read.spss("C:/Users/Nicola/Desktop/Relevant projects activities ACF/BRACED Final Evaluation/Evaluations/CSI_compil_2017.sav"), file="from_sav_data.csv", quote = FALSE, sep = ",")

Answer 3

import pandas as pd
import pyreadstat as py

df, meta = py.read_sav('file.SAV')
writer=pd.ExcelWriter ("file2.xlsx")
df.to_excel(writer, 'df')
df.to_csv('file2.csv')
writer.save()

Answer 4

Here is my class that finds all .sav files in subfolders and converts them to the .csv format. The class successfully coped with files up to 2 GB in size. If you only need to convert one file, just use the convert_sav_to_csv(file_path, dir_to_save=None) function. I used this answer a lot.

import glob
import pandas as pd
import os
import enum
import errno
import pyreadstat


# Enum for size units
class SizeUnit(enum.Enum):
    BYTES = 1
    KB = 2
    MB = 3
    GB = 4


class ConverterSavCsv:
    @staticmethod
    def convert_unit(size_in_bytes, unit):
        """ Convert the size from bytes to other units like KB, MB or GB"""
        if unit == SizeUnit.KB:
            return size_in_bytes / 1024
        elif unit == SizeUnit.MB:
            return size_in_bytes / (1024 * 1024)
        elif unit == SizeUnit.GB:
            return size_in_bytes / (1024 * 1024 * 1024)
        else:
            return size_in_bytes

    @staticmethod
    def get_file_size(file_name, size_type=SizeUnit.BYTES):
        """ Get file in size in given unit like KB, MB or GB"""
        size = os.path.getsize(file_name)
        return ConverterSavCsv.convert_unit(size, size_type)

    @staticmethod
    def find_all_sav_files(upper_dir_path):
        if not os.path.isdir(dir_path):
            raise NotADirectoryError(dir_path)
        files = glob.glob(upper_dir_path + '/**/*.sav', recursive=True)
        files_and_sizes = []
        for elem in files:
            elem_size = ConverterSavCsv.get_file_size(elem, SizeUnit.MB)
            files_and_sizes.append((elem, elem_size))
        result_df = pd.DataFrame(files_and_sizes, columns=['Path', 'Size'])
        result_df = result_df.sort_values(by=['Size'], ascending=False)
        result_df.to_excel(os.path.join(upper_dir_path, r'spss_files_sizes.xlsx'), index=False)
        return result_df

    @staticmethod
    def _dir_filepath_prepare_before_conversion(file_path, dir_to_save=None):
        if not os.path.isfile(file_path):
            raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), file_path)
        base_filename = os.path.basename(file_path)
        base_filename = os.path.splitext(base_filename)[0]
        if dir_to_save is None:
            current_dir = os.path.dirname(file_path)
            dir_to_save = os.path.join(current_dir, r'ready_csv')
            os.makedirs(dir_to_save)
        path_to_save = os.path.join(dir_to_save, base_filename+r'.csv')
        return path_to_save

    @staticmethod
    def convert_sav_to_csv(file_path, dir_to_save=None):
        path_to_save = ConverterSavCsv._dir_filepath_prepare_before_conversion(file_path, dir_to_save)
        _, meta_start = pyreadstat.read_sav(file_path, metadataonly=True)
        number_of_rows = meta_start.number_rows
        reader = pyreadstat.read_file_in_chunks(pyreadstat.read_sav, file_path, chunksize=1000)
        cnt = 0
        for df, meta in reader:
            if number_of_rows is not None:
                print('Processing rows', cnt * 1000, '-', (cnt + 1) * 1000, 'of', number_of_rows)
            if cnt > 0:
                write_mode = "a"
                header = False
            else:
                write_mode = "w"
                header = True
            df.to_csv(path_to_save, mode=write_mode, header=header)
            cnt += 1
        return ConverterSavCsv.get_file_size(path_to_save, SizeUnit.MB)

    @staticmethod
    def convert_all_sav_files_to_csv(sav_files_df, dir_to_save=None):
        if sav_files_df.empty \
                or 'Path' not in sav_files_df.columns \
                or 'Size' not in sav_files_df.columns:
            raise ValueError('Problems with input Pandas dataframe in convert_all_sav_files_to_csv function')
        if dir_to_save is not None:
            if not os.path.isdir(dir_to_save):
                os.makedirs(dir_to_save)
                # raise NotADirectoryError(errno.ENOENT, os.strerror(errno.ENOENT), dir_to_save)
        counter = 1
        files_total = sav_files_df.shape[0]
        for index, row in sav_files_df.iterrows():
            base_filename = os.path.basename(row['Path'])
            base_filename = os.path.splitext(base_filename)[0]
            print('Converting file', base_filename, '-', counter, 'of', files_total, ', with size', row['Size'], 'MB')
            ConverterSavCsv.convert_sav_to_csv(row['Path'], dir_to_save)
            counter = counter + 1
            print()
        return files_total


if __name__ == "__main__":
    dir_path = r'C:\Job\sav_files'
    dir_to_save_path = r'C:\Job\sav_files\csv_ready_files'
    # if dir_to_save_path is None,
    # it creates subdirectory ready_csv and saves there

    # find all SAV files in the directory and convert them to CSV
    converter = ConverterSavCsv()
    df_sav_files_path_size = converter.find_all_sav_files(dir_path)
    num_of_files = converter.convert_all_sav_files_to_csv(df_sav_files_path_size, dir_to_save_path)
    print(num_of_files, '.sav files converted to .csv and saved')

    # for single file
    # use function convert_sav_to_csv(file_path, dir_to_save=None)