def pcr_etf_option(dt_start, dt_end, name_code, df_res):
optionMkt = admin.table_options_mktdata()
Index_mkt = admin.table_indexes_mktdata()
query_pcr = admin.session_gc().query(optionMkt.c.dt_date, optionMkt.c.cd_option_type,
optionMkt.c.id_underlying,
func.sum(optionMkt.c.amt_holding_volume).label('total_holding_volume'),
func.sum(optionMkt.c.amt_trading_volume).label('total_trading_volume')
) \
.filter(optionMkt.c.dt_date >= dt_start) \
.filter(optionMkt.c.dt_date <= dt_end) \
.filter(optionMkt.c.name_code == name_code) \
.group_by(optionMkt.c.cd_option_type, optionMkt.c.dt_date, optionMkt.c.id_underlying)
df_pcr = pd.read_sql(query_pcr.statement, query_pcr.session.bind)
query_etf = admin.session_gc().query(Index_mkt.c.dt_date, Index_mkt.c.amt_close, Index_mkt.c.amt_open,
Index_mkt.c.id_instrument.label(c.Util.ID_UNDERLYING)) \
.filter(Index_mkt.c.dt_date >= dt_start).filter(Index_mkt.c.dt_date <= dt_end) \
.filter(Index_mkt.c.id_instrument == 'index_50etf')
df_50etf = pd.read_sql(query_etf.statement, query_etf.session.bind)
df = df_pcr.groupby(['dt_date', 'cd_option_type'
])['total_holding_volume',
'total_trading_volume'].sum().reset_index()
df_call = df[df['cd_option_type'] == 'call'].reset_index()
df_put = df[df['cd_option_type'] == 'put'].reset_index()
pc_ratio = []
for idx, row in df_call.iterrows():
row_put = df_put[df_put['dt_date'] == row['dt_date']]
pcr_trading = row_put['total_trading_volume'].values[0] / row[
'total_trading_volume']
pcr_holding = row_put['total_holding_volume'].values[0] / row[
'total_holding_volume']
pc_ratio.append({
'dt_date': row['dt_date'],
'tv_c': row['total_trading_volume'],
'tv_p': row_put['total_trading_volume'].values[0],
'hv_c': row['total_holding_volume'],
'hv_p': row_put['total_holding_volume'].values[0],
'tv_pcr': pcr_trading,
'hv_pcr': pcr_holding,
})
df_pcr = pd.DataFrame(pc_ratio)
df_pcr = pd.merge(df_pcr,
df_50etf[['dt_date', 'amt_close']],
how='left',
on=['dt_date'],
suffixes=['', '_r'])
df_pcr = df_pcr.sort_values(by='dt_date',
ascending=False).reset_index(drop=True)
df_res.loc[:, name_code + ':A:date'] = df_pcr['dt_date']
df_res.loc[:, name_code + ':B:tv_c'] = df_pcr['tv_c']
df_res.loc[:, name_code + ':C:tv_p'] = df_pcr['tv_p']
df_res.loc[:, name_code + ':E:date'] = df_pcr['dt_date']
df_res.loc[:, name_code + ':F:hv_c'] = df_pcr['hv_c']
df_res.loc[:, name_code + ':G:hv_p'] = df_pcr['hv_p']
df_res.loc[:, name_code + ':I:date'] = df_pcr['dt_date']
df_res.loc[:, name_code + ':J:tv_pcr'] = df_pcr['tv_pcr']
df_res.loc[:, name_code + ':K:hv_pcr'] = df_pcr['hv_pcr']
df_res.loc[:, name_code + ':L:amt_close'] = df_pcr['amt_close']
return df_res
def pcr(dt_start, dt_end, name_code, df_res):
optionMkt = admin.table_options_mktdata()
futureMkt = admin.table_futures_mktdata()
query_pcr = admin.session_mktdata().query(optionMkt.c.dt_date, optionMkt.c.cd_option_type,
optionMkt.c.id_underlying,
func.sum(optionMkt.c.amt_holding_volume).label('total_holding_volume'),
func.sum(optionMkt.c.amt_trading_volume).label('total_trading_volume')
) \
.filter(optionMkt.c.dt_date >= dt_start) \
.filter(optionMkt.c.dt_date <= dt_end) \
.filter(optionMkt.c.name_code == name_code) \
.group_by(optionMkt.c.cd_option_type, optionMkt.c.dt_date, optionMkt.c.id_underlying)
query_srf = admin.session_mktdata().query(futureMkt.c.dt_date, futureMkt.c.id_instrument,
futureMkt.c.amt_close, futureMkt.c.amt_trading_volume,
futureMkt.c.amt_settlement) \
.filter(futureMkt.c.dt_date >= dt_start) \
.filter(futureMkt.c.dt_date <= dt_end) \
.filter(futureMkt.c.name_code == name_code) \
.filter(futureMkt.c.flag_night != 1)
df_pcr = pd.read_sql(query_pcr.statement, query_pcr.session.bind)
df_srf = pd.read_sql(query_srf.statement, query_srf.session.bind)
# 按期权合约持仓量最大选取主力合约
df = df_pcr[df_pcr.groupby(['dt_date', 'cd_option_type'])['total_holding_volume'].transform(max) == df_pcr[
'total_holding_volume']]
df_call = df[df['cd_option_type'] == 'call'].reset_index()
df_put = df[df['cd_option_type'] == 'put'].reset_index()
pc_ratio = []
for idx, row in df_call.iterrows():
row_put = df_put[df_put['dt_date'] == row['dt_date']]
pcr_trading = row_put['total_trading_volume'].values[0] / row['total_trading_volume']
pcr_holding = row_put['total_holding_volume'].values[0] / row['total_holding_volume']
pc_ratio.append({'dt_date': row['dt_date'],
'tv_c': row['total_trading_volume'],
'tv_p': row_put['total_trading_volume'].values[0],
'hv_c': row['total_holding_volume'],
'hv_p': row_put['total_holding_volume'].values[0],
'tv_pcr': pcr_trading,
'hv_pcr': pcr_holding,
'id_instrument': row['id_underlying'],
})
df_pcr = pd.DataFrame(pc_ratio)
df_pcr = pd.merge(df_pcr, df_srf[['dt_date', 'id_instrument', 'amt_settlement']], how='left',
on=['dt_date', 'id_instrument'],
suffixes=['', '_r'])
df_pcr = df_pcr.sort_values(by='dt_date', ascending=False).reset_index(drop=True)
df_res.loc[:, 'A:date'] = df_pcr['dt_date']
df_res.loc[:, 'B:tv_c'] = df_pcr['tv_c']
df_res.loc[:, 'C:tv_p'] = df_pcr['tv_p']
df_res.loc[:, 'D'] = None
df_res.loc[:, 'E:date'] = df_pcr['dt_date']
df_res.loc[:, 'F:hv_c'] = df_pcr['hv_c']
df_res.loc[:, 'G:hv_p'] = df_pcr['hv_p']
df_res.loc[:, 'H'] = None
df_res.loc[:, 'I:date'] = df_pcr['dt_date']
df_res.loc[:, 'J:tv_pcr'] = df_pcr['tv_pcr']
df_res.loc[:, 'K:hv_pcr'] = df_pcr['hv_pcr']
df_res.loc[:, 'L:amt_settlement'] = df_pcr['amt_settlement']
df_res.loc[:, 'M'] = None
return df_res
def df_iv_at_the_money(dt_date, dt_start, namecode, df_srf):
optionMetrics = dbt.OptionMetrics
query_sro = admin.session_metrics().query(optionMetrics.dt_date, optionMetrics.id_instrument,
optionMetrics.id_underlying,
optionMetrics.amt_strike,
optionMetrics.cd_option_type, optionMetrics.pct_implied_vol) \
.filter(optionMetrics.dt_date >= dt_start) \
.filter(optionMetrics.dt_date <= dt_date)\
.filter(optionMetrics.name_code == namecode)
df_sro = pd.read_sql(query_sro.statement, query_sro.session.bind)
dates = df_sro['dt_date'].unique()
dict_iv_call = []
dict_iv_put = []
for date in dates:
df_volume_groupby = get_volume_groupby_id_option(admin.table_options_mktdata(), namecode, dt_start=date,
dt_end=date). \
sort_values(by='total_trading_volume', ascending=False).reset_index(drop=True)
id_c1 = df_volume_groupby.loc[0, 'id_underlying']
df0 = df_sro[(df_sro['dt_date'] == date)
& (df_sro['id_underlying'] == id_c1)]
df1 = df0[(df0['cd_option_type'] == 'call')]
amt_settle = \
df_srf[(df_srf['dt_date'] == date) & (df_srf['id_instrument'] == id_c1)]['amt_settlement'].values[0]
df1['diff'] = abs(df1['amt_strike'] - amt_settle)
df1 = df1.sort_values(by='diff', ascending=True)
k = df1.iloc[0]['amt_strike']
iv_call_c1 = df1.iloc[0]['pct_implied_vol'] * 100
iv_put_c1 = df0[(df0['cd_option_type'] == 'put') & (
df0['amt_strike'] == k)]['pct_implied_vol'].values[0] * 100
id_c2 = df_volume_groupby.loc[1, 'id_underlying']
df0 = df_sro[(df_sro['dt_date'] == date)
& (df_sro['id_underlying'] == id_c2)]
df1 = df0[(df0['cd_option_type'] == 'call')]
amt_settle = \
df_srf[(df_srf['dt_date'] == date) & (df_srf['id_instrument'] == id_c1)]['amt_settlement'].values[0]
df1['diff'] = abs(df1['amt_strike'] - amt_settle)
df1 = df1.sort_values(by='diff', ascending=True)
k = df1.iloc[0]['amt_strike']
iv_call_c2 = df1.iloc[0]['pct_implied_vol'] * 100
iv_put_c2 = df0[(df0['cd_option_type'] == 'put') & (
df0['amt_strike'] == k)]['pct_implied_vol'].values[0] * 100
dict_iv_call.append({
'dt_date': date,
'iv_c1': iv_call_c1,
'iv_c2': iv_call_c2,
'underlying_c1': id_c1,
'underlying_c2': id_c2,
})
dict_iv_put.append({
'dt_date': date,
'iv_c1': iv_put_c1,
'iv_c2': iv_put_c2,
'underlying_c1': id_c1,
'underlying_c2': id_c2,
})
df_call = pd.DataFrame(dict_iv_call)
df_put = pd.DataFrame(dict_iv_put)
return df_call, df_put
def commodity_option_market_overview_by_month(start_date, end_date, name_code):
optionMkt = admin.table_options_mktdata()
futureMkt = admin.table_futures_mktdata()
query = admin.session_gc().query(optionMkt.c.dt_date,optionMkt.c.id_underlying,
func.sum(optionMkt.c.amt_trading_volume).label('option_trading_volume'),
func.sum(optionMkt.c.amt_trading_value).label('option_trading_value')
) \
.filter(optionMkt.c.dt_date >= start_date) \
.filter(optionMkt.c.dt_date <= end_date) \
.filter(optionMkt.c.name_code == name_code) \
.group_by(optionMkt.c.dt_date, optionMkt.c.id_underlying)
df_option_trading = pd.read_sql(query.statement, query.session.bind)
query_future = admin.session_gc().query(futureMkt.c.dt_date,futureMkt.c.id_instrument,
func.sum(futureMkt.c.amt_trading_volume).label('future_trading_volume')
) \
.filter(futureMkt.c.dt_date >= start_date) \
.filter(futureMkt.c.dt_date <= end_date) \
.filter(futureMkt.c.name_code == name_code) \
.group_by(futureMkt.c.dt_date, futureMkt.c.id_instrument)
df_future_trading = pd.read_sql(query_future.statement,
query_future.session.bind)
query_option_holding = admin.session_gc().query(optionMkt.c.dt_date, optionMkt.c.id_underlying,
func.sum(optionMkt.c.amt_holding_volume).label('option_holding_volume')) \
.filter(optionMkt.c.dt_date >= start_date) \
.filter(optionMkt.c.dt_date <= end_date) \
.filter(optionMkt.c.name_code == name_code)\
.group_by(optionMkt.c.dt_date, optionMkt.c.id_underlying) #每日日盘收盘持仓数据
df_option_holding = pd.read_sql(query_option_holding.statement,
query_option_holding.session.bind)
query_future_holding = admin.session_gc().query(futureMkt.c.dt_date,futureMkt.c.id_instrument,
func.sum(futureMkt.c.amt_holding_volume).label('future_holding_volume')) \
.filter(futureMkt.c.dt_date >= start_date) \
.filter(futureMkt.c.dt_date <= end_date) \
.filter(futureMkt.c.name_code == name_code) \
.group_by(futureMkt.c.dt_date, futureMkt.c.id_instrument) #每日日盘收盘持仓数据
df_future_holding = pd.read_sql(query_future_holding.statement,
query_future_holding.session.bind)
# new_df = pd.merge(A_df, B_df, how='left', left_on=['A_c1', 'c2'], right_on=['B_c1', 'c2'])
df_future = pd.merge(df_future_holding,
df_future_trading,
on=[c.Util.DT_DATE, c.Util.ID_INSTRUMENT])
df_option = pd.merge(df_option_holding,
df_option_trading,
on=[c.Util.DT_DATE, c.Util.ID_UNDERLYING])
df = pd.merge(df_option,
df_future,
left_on=[c.Util.DT_DATE, c.Util.ID_UNDERLYING],
right_on=[c.Util.DT_DATE, c.Util.ID_INSTRUMENT])
# df = df.groupby([c.Util.DT_DATE,c.Util.ID_UNDERLYING,c.Util.ID_INSTRUMENT])['col3'].sum()
return df
def commodity_option_market_overview(start_date, end_date, name_code):
optionMkt = admin.table_options_mktdata()
futureMkt = admin.table_futures_mktdata()
query = admin.session_gc().query(optionMkt.c.dt_date,
func.sum(optionMkt.c.amt_trading_volume).label('option_trading_volume'),
func.sum(optionMkt.c.amt_trading_value).label('option_trading_value')
) \
.filter(optionMkt.c.dt_date >= start_date) \
.filter(optionMkt.c.dt_date <= end_date) \
.filter(optionMkt.c.name_code == name_code) \
.group_by(optionMkt.c.dt_date)
df_option_trading = pd.read_sql(query.statement, query.session.bind)
query_future = admin.session_gc().query(futureMkt.c.dt_date,
func.sum(futureMkt.c.amt_trading_volume).label('future_trading_volume')
) \
.filter(futureMkt.c.dt_date >= start_date) \
.filter(futureMkt.c.dt_date <= end_date) \
.filter(futureMkt.c.name_code == name_code) \
.group_by(futureMkt.c.dt_date)
df_future_trading = pd.read_sql(query_future.statement,
query_future.session.bind)
query_option_holding = admin.session_gc().query(optionMkt.c.dt_date,
func.sum(optionMkt.c.amt_holding_volume).label('option_holding_volume')) \
.filter(optionMkt.c.dt_date >= start_date) \
.filter(optionMkt.c.dt_date <= end_date) \
.filter(optionMkt.c.name_code == name_code) \
.group_by(optionMkt.c.dt_date) #每日日盘收盘持仓数据
df_option_holding = pd.read_sql(query_option_holding.statement,
query_option_holding.session.bind)
query_future_holding = admin.session_gc().query(futureMkt.c.dt_date,
func.sum(futureMkt.c.amt_holding_volume).label('future_holding_volume')) \
.filter(futureMkt.c.dt_date >= start_date) \
.filter(futureMkt.c.dt_date <= end_date) \
.filter(futureMkt.c.name_code == name_code) \
.group_by(futureMkt.c.dt_date) #每日日盘收盘持仓数据
df_future_holding = pd.read_sql(query_future_holding.statement,
query_future_holding.session.bind)
# df = pd.merge(df_option_trading,df_future_trading[[c.Util.DT_DATE,'future_trading_volume']],on=c.Util.DT_DATE)
# df = pd.merge(df,df_option_holding,on=c.Util.DT_DATE)
# df = pd.merge(df,df_future_holding,on=c.Util.DT_DATE)
df_future = pd.merge(df_future_holding,
df_future_trading,
on=[c.Util.DT_DATE])
df_option = pd.merge(df_option_holding,
df_option_trading,
on=[c.Util.DT_DATE])
df = pd.merge(df_option, df_future, on=[c.Util.DT_DATE])
return df
def get_comoption_mktdata(start_date, end_date, name_code):
# Future_mkt = dbt.FutureMkt
table_mktdata = admin.table_futures_mktdata_gc()
Option_mkt = admin.table_options_mktdata()
options = dbt.Options
query_mkt = admin.session_gc(). \
query(Option_mkt.c.dt_date, Option_mkt.c.id_instrument, Option_mkt.c.id_underlying,
Option_mkt.c.code_instrument, Option_mkt.c.amt_close, Option_mkt.c.amt_open,
Option_mkt.c.amt_settlement,
Option_mkt.c.amt_last_settlement, Option_mkt.c.amt_trading_volume,Option_mkt.c.amt_trading_value,
Option_mkt.c.pct_implied_vol, Option_mkt.c.amt_holding_volume,
Option_mkt.c.amt_trading_volume
) \
.filter(Option_mkt.c.dt_date >= start_date).filter(Option_mkt.c.dt_date <= end_date) \
.filter(Option_mkt.c.name_code == name_code).filter(Option_mkt.c.flag_night != 1)
query_option = admin.session_mktdata(). \
query(options.id_instrument, options.cd_option_type, options.amt_strike, options.name_contract_month,
options.dt_maturity, options.nbr_multiplier) \
.filter(and_(options.dt_listed <= end_date, options.dt_maturity >= start_date))
query_srf = admin.session_gc(). \
query(table_mktdata.c.dt_date,
table_mktdata.c.id_instrument.label(c.Util.ID_UNDERLYING),
table_mktdata.c.amt_settlement.label(c.Util.AMT_UNDERLYING_CLOSE),
table_mktdata.c.amt_open.label(c.Util.AMT_UNDERLYING_OPEN_PRICE)) \
.filter(table_mktdata.c.dt_date >= start_date).filter(table_mktdata.c.dt_date <= end_date) \
.filter(table_mktdata.c.name_code == name_code)
df_srf = pd.read_sql(query_srf.statement, query_srf.session.bind)
df_mkt = pd.read_sql(query_mkt.statement, query_mkt.session.bind)
df_contract = pd.read_sql(query_option.statement,
query_option.session.bind)
df_option = df_mkt.join(df_contract.set_index('id_instrument'),
how='left',
on='id_instrument')
df_option_metrics = pd.merge(df_option,
df_srf,
how='left',
on=['dt_date', 'id_underlying'],
suffixes=['', '_r'])
return df_option_metrics
def get_50option_mktdata(start_date, end_date):
Index_mkt = dbt.IndexMkt
Option_mkt = admin.table_options_mktdata()
options = dbt.Options
util = c.Util
query_mkt = admin.session_gc().query(Option_mkt.c.dt_date, Option_mkt.c.id_instrument,
Option_mkt.c.code_instrument,
Option_mkt.c.amt_open,
Option_mkt.c.amt_close, Option_mkt.c.amt_settlement,
Option_mkt.c.amt_last_settlement,Option_mkt.c.amt_trading_value,
Option_mkt.c.amt_trading_volume, Option_mkt.c.amt_holding_volume,
Option_mkt.c.pct_implied_vol
) \
.filter(Option_mkt.c.dt_date >= start_date).filter(Option_mkt.c.dt_date <= end_date) \
.filter(Option_mkt.c.datasource == 'wind').filter(Option_mkt.c.name_code == '50etf')
query_option = admin.session_mktdata().query(options.id_instrument, options.cd_option_type,
options.amt_strike, options.name_contract_month,
options.dt_maturity, options.nbr_multiplier) \
.filter(and_(options.dt_listed <= end_date, options.dt_maturity >= start_date))
query_etf = admin.session_gc().query(Index_mkt.dt_date, Index_mkt.amt_close, Index_mkt.amt_open,
Index_mkt.id_instrument.label(util.ID_UNDERLYING)) \
.filter(Index_mkt.dt_date >= start_date).filter(Index_mkt.dt_date <= end_date) \
.filter(Index_mkt.id_instrument == 'index_50etf')
df_mkt = pd.read_sql(query_mkt.statement, query_mkt.session.bind)
df_contract = pd.read_sql(query_option.statement,
query_option.session.bind)
df_50etf = pd.read_sql(query_etf.statement, query_etf.session.bind).rename(
columns={
'amt_close': util.AMT_UNDERLYING_CLOSE,
'amt_open': util.AMT_UNDERLYING_OPEN_PRICE
})
df_option = df_mkt.join(df_contract.set_index('id_instrument'),
how='left',
on='id_instrument')
df_option_metrics = df_option.join(df_50etf.set_index('dt_date'),
how='left',
on='dt_date')
return df_option_metrics
def trade_volume(dt_date, dt_last_week, w, nameCode, core_instrumentid):
pu = PlotUtil()
options_mkt = admin.table_options_mktdata()
evalDate = dt_date.strftime("%Y-%m-%d") # Set as Friday
plt.rcParams['font.sans-serif'] = ['STKaiti']
plt.rcParams.update({'font.size': 15})
"""当日成交持仓量 """
query_volume = admin.session_mktdata().query(options_mkt.c.dt_date,
options_mkt.c.cd_option_type,
options_mkt.c.amt_strike,
options_mkt.c.amt_holding_volume,
options_mkt.c.amt_trading_volume,
options_mkt.c.amt_close,
options_mkt.c.pct_implied_vol
) \
.filter(or_(options_mkt.c.dt_date == evalDate,options_mkt.c.dt_date == dt_last_week)) \
.filter(options_mkt.c.id_underlying == core_instrumentid)\
.filter(options_mkt.c.flag_night != 1)
df_2d = pd.read_sql(query_volume.statement, query_volume.session.bind)
df = df_2d[df_2d['dt_date'] == dt_date].reset_index()
df_lw = df_2d[df_2d['dt_date'] == dt_last_week].reset_index()
df_call = df[df['cd_option_type'] == 'call'].reset_index()
df_put = df[df['cd_option_type'] == 'put'].reset_index()
dflw_call = df_lw[df_lw['cd_option_type'] == 'call'].reset_index()
dflw_put = df_lw[df_lw['cd_option_type'] == 'put'].reset_index()
call_deltas = []
put_deltas = []
for idx, row in df_call.iterrows():
row_put = df_put.loc[idx]
strike = row['amt_strike']
rowlw_call = dflw_call[dflw_call['amt_strike'] == strike]
rowlw_put = dflw_put[dflw_put['amt_strike'] == strike]
last_holding_call = 0.0
last_holding_put = 0.0
try:
last_holding_call = rowlw_call['amt_holding_volume'].values[0]
except:
pass
try:
last_holding_put = rowlw_put['amt_holding_volume'].values[0]
except:
pass
call_delta = row['amt_holding_volume'] - last_holding_call
put_delta = row_put['amt_holding_volume'] - last_holding_put
call_deltas.append(call_delta)
put_deltas.append(put_delta)
if nameCode == 'sr':
wt = 25
else:
wt = 15
strikes = df_call['amt_strike'].tolist()
strikes1 = df_call['amt_strike'] + wt
holding_call = df_call['amt_holding_volume'].tolist()
holding_put = df_put['amt_holding_volume'].tolist()
trading_call = df_call['amt_trading_volume'].tolist()
trading_put = df_put['amt_trading_volume'].tolist()
df_results = pd.DataFrame({
'0 call iv':
df_call['pct_implied_vol'].tolist(),
'1 call delta_holding':
call_deltas,
'2 call holding':
df_call['amt_holding_volume'].tolist(),
'3 call trading':
df_call['amt_trading_volume'].tolist(),
'4 call price':
df_call['amt_close'].tolist(),
'5 strikes':
df_put['amt_strike'].tolist(),
'6 put price':
df_put['amt_close'].tolist(),
'7 put trading':
df_put['amt_trading_volume'].tolist(),
'8 put holding':
df_put['amt_holding_volume'].tolist(),
'9 put delta_holding':
put_deltas,
'91 put iv':
df_put['pct_implied_vol'].tolist()
})
df_results.to_csv('../data/' + nameCode + '_holdings_' + evalDate + '.csv')
ldgs = ['持仓量（看涨）', '持仓量（看跌）', '成交量（看涨）', '成交量（看跌）']
f3, ax3 = plt.subplots()
p1 = ax3.bar(strikes, holding_call, width=wt, color=pu.colors[0])
p2 = ax3.bar(strikes1, holding_put, width=wt, color=pu.colors[1])
p3, = ax3.plot(strikes,
trading_call,
color=pu.colors[2],
linestyle=pu.lines[2],
linewidth=2)
p4, = ax3.plot(strikes,
trading_put,
color=pu.colors[3],
linestyle=pu.lines[3],
linewidth=2)
ax3.legend([p1, p2, p3, p4],
ldgs,
bbox_to_anchor=(0., 1.02, 1., .102),
loc=3,
ncol=4,
mode="expand",
borderaxespad=0.,
frameon=False)
ax3.spines['top'].set_visible(False)
ax3.spines['right'].set_visible(False)
ax3.yaxis.set_ticks_position('left')
ax3.xaxis.set_ticks_position('bottom')
f3.set_size_inches((12, 8))
f3.savefig('../data/' + nameCode + '_holdings_' + evalDate + '.png',
dpi=300,
format='png',
bbox_inches='tight')