from tableausdk import Extract as tde # **Module change from py2 for py3
from tableausdk import Types as tdeTypes # **
import os
#Step 1: Create the Extract File
if os.path.isfile('trivialExample.tde'):
os.remove('trivialExample.tde')
tdefile = tde.Extract('trivialExample.tde')
#Step 2: Create the tableDef
tableDef = tde.TableDefinition()
tableDef.addColumn('rowID', tdeTypes.Type.CHAR_STRING) # Column index 0. **
tableDef.addColumn('value', tdeTypes.Type.DOUBLE) #Column index 1. **
#Step 3: Create the table in the image of the tableDef
table = tdefile.addTable('Extract', tableDef)
#Step 4: Create some rows and insert them into the table
newrow = tde.Row(tableDef)
for i in range(1, 101):
newrow.setCharString(0, 'Row ' + str(i))
newrow.setDouble(1, i)
table.insert(newrow)
#Step 5: Close the tde
tdefile.close()
def build(self):
# Set filename
tdefileName = self.tde_settings_ins.tde_file
if os.path.isfile(tdefileName):
os.remove(tdefileName)
#set the extract object instance
tdefile = tdeEx.Extract(tdefileName)
# create the column schema
tableDef = tdeEx.TableDefinition()
columnsListedInOrder = []
outputColIndex = {}
outputColNumber = 0
for columnName, index in sorted(
self.tde_settings_ins.columnIndex.items(), key=lambda x: x[1]):
BasicDataType = self.tde_settings_ins.columns[columnName]
tableDef.addColumn(columnName,
tdeSettings.schemaIniTypeMap[BasicDataType])
outputColIndex[columnName] = outputColNumber
outputColNumber += 1
#add the column schema to the ojbect instance of EXTRACT
table = tdefile.addTable("Extract", tableDef)
rowNo = 0
rowsIterator = self.tde_settings_ins.inputInfo.fileInformation.yieldRowsBase(
)
printedBool = False
startTime = datetime.datetime.now()
print "Build start time:", datetime.datetime.strftime(
startTime, '%Y-%m-%d %H:%M:%S')
prevPct = -1
for myReaderRow in rowsIterator:
putRow = tdeEx.Row(tableDef)
for columnName, index in self.tde_settings_ins.columnIndex.items():
#if not column headers need to change this to index of the column
if (tdeSettings.schemaIniTypeMap[
self.tde_settings_ins.columns[columnName]] !=
tde.Type.UNICODE_STRING and myReaderRow[index] == ""):
putRow.setNull(outputColIndex[columnName])
else:
try:
#valueDecoded=myReaderRow[index].decode(self.tde_settings_ins.inputInfo.fileInformation.encodeing)
self.fieldSetterMap[tdeSettings.schemaIniTypeMap[
self.tde_settings_ins.columns[columnName]]](
putRow, outputColIndex[columnName],
myReaderRow[index])
except:
print "column name", columnName, "contains invalid data"
print "value: ", myReaderRow[index]
print sys.exc_info()[0]
raise
table.insert(putRow)
# Output progress line
#only print percent one time when we hit that percent
rowNo += 1
pct = math.floor(
self.tde_settings_ins.inputInfo.fileInformation.getPercent(
rowNo))
# print "row:",rowNo
# print "pct:",pct
# print "value: ",pct %2
if pct % 2 == 0:
if printedBool == False or prevPct != pct:
print "Percent:", round(pct, 0), "%, row:", rowNo
printedBool = True
prevPct = pct
else:
printedBool = False
print "tde file built to:", tdefileName
endTime = datetime.datetime.now()
print "End Time:", endTime
diff = endTime - startTime
print "Build elapsed time:", diff
def create(target_filepath, dataframe):
column = []
## 宣告Dataframe(左)pandas資料格式 對應 到TDE的資料格式(右)。
fieldMap = {
'float64': tde.Types.Type.DOUBLE,
'float32': tde.Types.Type.DOUBLE,
'int64': tde.Types.Type.INTEGER,
'int8': tde.Types.Type.INTEGER,
'object': tde.Types.Type.UNICODE_STRING,
'bool': tde.Types.Type.DOUBLE,
'datetime64[ns]': tde.Types.Type.DATETIME
}
## 若TDE已存在，則刪除。
if os.path.isfile(target_filepath):
os.remove(target_filepath)
tdefile = tde.Extract(target_filepath)
schema = tde.TableDefinition() # define the table definition
colnames = dataframe.columns # dataframe all column name
coltypes = dataframe.dtypes # dataframe all column's datatype
## 根據Dataframe建立Tde Schema
for i in range(0, len(colnames)):
cname = colnames[i]
ctype = fieldMap.get(str(coltypes[i]))
schema.addColumn(cname, ctype) # 加入一行欄位(名稱，資料格式)
column.append(cname)
## 開啟TDE FILE寫入資料
with tdefile as extract:
table = extract.addTable("Extract", schema)
for r in range(0, dataframe.shape[0]):
row = tde.Row(schema)
## Loop 顯示剩餘筆數
count = dataframe.shape[0] - r - 1
print('剩' + str(count) + '筆', end="\r")
## 每筆資料依照格式寫入TDE內
for c in range(0, len(coltypes)):
if str(coltypes[c]) == 'float64':
if dataframe.iloc[r, c] is None:
row.setNull(c)
else:
row.setDouble(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'float32':
row.setDouble(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'int64':
if dataframe.iloc[r, c] is None:
row.setNull(c)
else:
row.setInteger(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'int8':
row.setInteger(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'object':
if dataframe.iloc[r, c] is None:
row.setNull(c)
else:
row.setString(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'bool':
row.setBoolean(c, dataframe.iloc[r, c])
elif str(coltypes[c]) == 'datetime64[ns]':
try:
dt = dataframe.iloc[r, c]
row.setDateTime(c, dt.year, dt.month, dt.day, dt.hour,
dt.minute, dt.second, 0)
except:
row.setNull(c)
else:
row.setNull(c)
## insert the row
table.insert(row)
## close the tdefile
tdefile.close()
def extract(file_name):
# move file to /extract
# cd to /extract
# if there is no extract called TRACK_TERM then create one, otherwise append to TRACK_TERM
# define data model for extract
global WORKING_DIRECTORY, TRACK_TERM
if not os.path.exists(WORKING_DIRECTORY + '/extract/'):
os.makedirs(WORKING_DIRECTORY + '/extract/')
from_path = WORKING_DIRECTORY + '/' + file_name
to_path = WORKING_DIRECTORY + '/extract/' + file_name
os.rename(from_path, to_path)
os.chdir(WORKING_DIRECTORY + '/extract')
# define the extract
with data.Extract(TRACK_TERM + '.tde') as extract:
tableDef = data.TableDefinition()
# define the columns and the data types in the extract
tableDef.addColumn('lang', types.Type.CHAR_STRING) #0
tableDef.addColumn('sentiment', types.Type.DOUBLE) #1
tableDef.addColumn('country', types.Type.CHAR_STRING) #2
tableDef.addColumn('created_at', types.Type.DATETIME) #3
tableDef.addColumn('tweet_text', types.Type.CHAR_STRING) #4
tableDef.addColumn('Longitude', types.Type.DOUBLE) #5
tableDef.addColumn('source', types.Type.CHAR_STRING) #6
tableDef.addColumn('user', types.Type.CHAR_STRING) #7
tableDef.addColumn('Latitude', types.Type.DOUBLE) #8
table = None
if not extract.hasTable('Extract'):
# Table does not exist, so create it.
print "Creating a new extract"
table = extract.addTable('Extract', tableDef)
else:
# Table exists, so append the new data.
print "Appending to an existing extract"
table = extract.openTable('Extract')
new_row = data.Row(tableDef)
# read the data from the CSV into the extract row object
with open(file_name, 'r') as inf:
reader = csv.DictReader(inf, delimiter=',', lineterminator='\n')
for row in reader:
# insert data into the row object in the correct order as defined above
new_row.setCharString(0, row['lang'])
sentiment = float(row['sentiment'])
new_row.setDouble(1, sentiment)
new_row.setCharString(2, row['country'])
# parse the twitter date string:
# Mon Sep 21 11:03:53 +0000 2015
# %a %b %d %H:%M:%S +0000 %Y
date_object = datetime.strptime(row['created_at'],
'%a %b %d %H:%M:%S +0000 %Y')
year = int(datetime.strftime(date_object, '%Y'))
month = int(datetime.strftime(date_object, '%m'))
day = int(datetime.strftime(date_object, '%d'))
hour = int(datetime.strftime(date_object, '%H'))
min = int(datetime.strftime(date_object, '%M'))
sec = int(datetime.strftime(date_object, '%S'))
frac = 0 # fractions of a second aka milliseconds
new_row.setDateTime(3, year, month, day, hour, min, sec, frac)
new_row.setCharString(4, row['tweet_text'])
# check if there is a value for longitude, otherwise write a 0
try:
longitude = float(row['longitude'])
except:
longitude = 0
new_row.setDouble(5, longitude)
new_row.setCharString(6, row['source'])
new_row.setCharString(7, row['user'])
# check if there is a value for latitude, otherwise write a 0
try:
latitude = float(row['latitude'])
except:
latitude = 0
new_row.setDouble(8, latitude)
table.insert(new_row)
# if the process fails we want to be able to re-run it without collisions between file names
# so give each file a unique name (unix time stamp in this case).
os.rename(file_name, str(time.time()).split('.')[0] + '.csv')
# cd back to working directory
os.chdir(WORKING_DIRECTORY)
return