我有一组图片功能（数据类型是python列表），如下所示，我使用pickle.dump列表到二进制文件并存储在表中。我想对功能进行多维数据集处理，但出现了错误，有没有一种方法可以对多维数据集（字节）进行处理。
[-0.07693052 0.13103333 0.03456296 -0.03198523 -0.11783121 -0.00864347
-0.08238359 -0.11123394 0.11118498 -0.12523678 0.17256702 -0.10489954
-0.21543087 -0.08727957 -0.07703612 0.24160105 -0.22249943 -0.15564784
-0.00786823 0.05911971 0.14252234 -0.00624161 0.01735243 0.0219292
-0.10636593 -0.33734819 -0.05135369 -0.07491996 -0.05320961 -0.02807039
-0.04177326 0.02587014 -0.16554965 -0.06286186 0.03911624 0.02000083
-0.02210705 -0.10491423 0.2021113 -0.00044226 -0.29604077 0.01300796
0.07944702 0.2044373 0.13803098 0.00925212 0.03331146 -0.1774627
0.12463985 -0.2000913 -0.00649202 0.13762167 0.08091226 0.07051353
0.00342046 -0.10343951 0.04754291 0.10721903 -0.11586423 -0.03075701
0.09388214 -0.03318139 0.01716307 -0.12743376 0.25524437 0.04930455
-0.09589496 -0.20655566 0.10854267 -0.08835522 -0.1341148 0.07985611
-0.10606824 -0.17955625 -0.26543319 -0.03198857 0.34081709 0.10754494
-0.19059491 0.03278148 -0.07174835 -0.04601264 0.08218642 0.20006074
0.02503737 0.06809284 -0.02955326 -0.03687159 0.25551286 -0.1125199
-0.01698952 0.24861689 0.03379418 0.05015399 0.00976648 -0.02963495
-0.07006045 0.08782617 -0.12364717 -0.03777558 0.04694367 -0.01261023
0.00260606 0.13151917 -0.17752528 0.1014059 -0.04664776 0.02771331
0.03564191 -0.02253933 -0.0823142 -0.05387688 0.11149541 -0.19149995
0.16104887 0.15177445 0.09675844 0.07879145 0.14599283 0.07733928
0.01200367 -0.01283191 -0.23660482 -0.01472425 0.1749115 -0.04009932
0.06074818 -0.01628037]
以下是外壳显示的方式：
face_service_dev=# select cube(feature1) from device_devicephotomodel;
ERROR: function cube(bytea) does not exist
LINE 1: select cube(feature1) from device_devicephotomodel;
^
HINT: No function matches the given name and argument types. You might need to add explicit type casts.
和表的ddl如下：
CREATE TABLE public.device_devicephotomodel
(
id bigint NOT NULL DEFAULT nextval('device_devicephotomodel_id_seq'::regclass),
upload_time timestamp with time zone NOT NULL,
take_photo_time timestamp with time zone NOT NULL,
path character varying(24) COLLATE pg_catalog."default" NOT NULL,
label integer NOT NULL,
feature1 bytea,
feature2 bytea,
cluster_id integer,
device_id integer,
并且我使用python psycopg2插入表，因为该功能我使用pickle.dumps（）转换为二进制。