search

sql-machine-learning / sqlflow

Brings SQL and AI together.
https://sqlflow.org
Apache License 2.0
5.07k stars 699 forks source link

SQLFlow error message examples #2165

Open Yancey1989 opened 4 years ago

Yancey1989 commented 4 years ago

SQLFlow Error Message Examples

This issue sorting some error cases according to the following bullets:

  1. reproduced SQL
  2. error message
  3. what we wanted(optional)

DataBase Connection Error

This error can occur on any SQL

Error Messages:

ERROR: runSQLProgram error: GetAlisaTask got a bad result, response={"returnCode":"11020295002","requestId":"0ab411db15880473516257764d0bda","returnMessage":"调用Alisa失败:java.net.SocketException: Connection reset","returnErrorSolution":""}

What we want:

Alisa database driver error: {"returnCode":"11020295002","requestId":"0ab411db15880473516257764d0bda","returnMessage":"调用Alisa失败:java.net.SocketException: Connection reset","returnErrorSolution":""}

PAI submitter program runtime error

SELECT * FROM alifin_jtest_dev.sqlflow_titanic_train
TO TRAIN DNNClassifier1 WITH
model.hidden_units=[200, 200, 200]
LABEL survived
INTO my_titanic_model;
runSQLProgram error: failed: pai -name tensorflow1150 -project 
****, detailed error message on URL: http://pai-logviwer-url

What we want:

Model Zoo check error: DNNClassifier1 does not exists, please check the
available model list on: http://model_zoo.sqlflow.org/list

Syntax Error

SELECT * FROM alifin_jtest_dev.sqlflow_titanic_train
TO TO TRAIN DNNClassifier WITH
model.hidden_units=[200, 200, 200]
model.no_exits=abc,
LABEL survived
INTO my_titanic_model;

Error Message:

runSQLProgram error: syntax error: at (3 ~ 5)-th runes near "TO TRAIN D"

What we want:

syntax error: ... TO TO TRAIN ..."

Attribute Check Error

SELECT * FROM alifin_jtest_dev.sqlflow_titanic_train
TO TRAIN DNNClassifier WITH
model.hidden_units=[200, 200, 200],
model.no_exits=abc
LABEL survived
INTO my_titanic_model;

Error Message:

runSQLProgram error: unsupported attribute model.no_exits

What we want:

attributed check error: unsupported attribute model.no_exits in DNNClassifier, allowed attributes of DNNClassifier please go to http://sqlflow.org/models/dnnclassifier
shendiaomo commented 4 years ago

There're many other errors about attribute checking or other semantics checking:

Type Checking Failed

SELECT * from iris.train TO TRAIN DNNClassifier WITH model.hidden_units=x LABEL class INTO iris.xgb;

What we have:

Traceback (most recent call last):

  File "<stdin>", line 136, in <module>

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train.py", line 126, in train

    save_checkpoints_steps, validation_metrics)

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 65, in estimator_train_and_save

    eval_start_delay_secs, eval_throttle_secs)

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 105, in estimator_train_compiled

    estimator.train(lambda: train_dataset_fn(), max_steps=train_max_steps)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 370, in train

    loss = self._train_model(input_fn, hooks, saving_listeners)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1160, in _train_model

    return self._train_model_default(input_fn, hooks, saving_listeners)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1190, in _train_model_default

    features, labels, ModeKeys.TRAIN, self.config)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1148, in _call_model_fn

    model_fn_results = self._model_fn(features=features, **kwargs)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 764, in _model_fn

    batch_norm=batch_norm)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 573, in dnn_model_fn_v2

    mode=mode)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 508, in _dnn_model_fn_builder_v2

    name='dnn')

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 328, in __init__

    name=hidden_shared_name)

  File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/layers/core.py", line 995, in __init__

    self.units = int(units)

ValueError: invalid literal for int() with base 10: 'x'

ERROR: rpc error: code = Unknown desc = runSQLProgram error: failed: exit status 1
==========Program==========

import copy
import tensorflow as tf
from sqlflow_submitter.tensorflow.train import train
from sqlflow_submitter.tensorflow.get_tf_version import tf_is_version2
from tensorflow.estimator import (DNNClassifier,
                                  DNNRegressor,
                                  LinearClassifier,
                                  LinearRegressor,
                                  BoostedTreesClassifier,
                                  BoostedTreesRegressor,
                                  DNNLinearCombinedClassifier,
                                  DNNLinearCombinedRegressor)
if tf_is_version2():
    from tensorflow.keras.optimizers import Adadelta, Adagrad, Adam, Adamax, Ftrl, Nadam, RMSprop, SGD
    from tensorflow.keras.losses import BinaryCrossentropy, CategoricalCrossentropy, CategoricalHinge, CosineSimilarity, Hinge, Huber, KLDivergence, LogCosh, MeanAbsoluteError, MeanAbsolutePercentageError, MeanSquaredError, MeanSquaredLogarithmicError, Poisson, SparseCategoricalCrossentropy, SquaredHinge
else:
    from tensorflow.train import AdadeltaOptimizer, AdagradOptimizer, AdamOptimizer, FtrlOptimizer, RMSPropOptimizer, GradientDescentOptimizer, MomentumOptimizer
    from tensorflow.keras.losses import BinaryCrossentropy, CategoricalCrossentropy, CategoricalHinge, CosineSimilarity, Hinge, Huber, KLDivergence, LogCosh, MeanAbsoluteError, MeanAbsolutePercentageError, MeanSquaredError, MeanSquaredLogarithmicError, Poisson, SparseCategoricalCrossentropy, SquaredHinge
try:
    import sqlflow_models
except Exception as e:
    print("failed to import sqlflow_models: %s", e)

feature_column_names = [
"sepal_length",

"sepal_width",

"petal_length",

"petal_width",
]

# feature_column_names_map is used to determine the order of feature columns of each target:
# e.g. when using DNNLinearCombinedClassifer.
# feature_column_names_map will be saved to a single file when using PAI.
feature_column_names_map = dict()

feature_column_names_map["feature_columns"] = ["sepal_length","sepal_width","petal_length","petal_width",]

feature_metas = dict()

feature_metas["sepal_length"] = {
    "feature_name": "sepal_length",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["sepal_width"] = {
    "feature_name": "sepal_width",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["petal_length"] = {
    "feature_name": "petal_length",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["petal_width"] = {
    "feature_name": "petal_width",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

label_meta = {
    "feature_name": "class",
    "dtype": "int64",
    "delimiter": "",
    "shape": [],
    "is_sparse": "false" == "true"
}

model_params=dict()

model_params["hidden_units"]="x"

# Construct optimizer objects to pass to model initializer.
# The original model_params is serializable (do not have tf.xxx objects).
model_params_constructed = copy.deepcopy(model_params)
for optimizer_arg in ["optimizer", "dnn_optimizer", "linear_optimizer"]:
    if optimizer_arg in model_params_constructed:
        model_params_constructed[optimizer_arg] = eval(model_params_constructed[optimizer_arg])

if "loss" in model_params_constructed:
    model_params_constructed["loss"] = eval(model_params_constructed["loss"])

# feature_columns_code will be used to save the training informations together
# with the saved model.
feature_columns_code = """{"feature_columns": [tf.feature_column.numeric_column("sepal_length", shape=[1]),
tf.feature_column.numeric_column("sepal_width", shape=[1]),
tf.feature_column.numeric_column("petal_length", shape=[1]),
tf.feature_column.numeric_column("petal_width", shape=[1])]}"""
feature_columns = eval(feature_columns_code)

train_max_steps = 0
train_max_steps = None if train_max_steps == 0 else train_max_steps

train(datasource="mysql://root:root@tcp(127.0.0.1:3306)/?maxAllowedPacket=0",
      estimator=DNNClassifier,
      select="""SELECT * from iris.train """,
      validation_select="""""",
      feature_columns=feature_columns,
      feature_column_names=feature_column_names,
      feature_metas=feature_metas,
      label_meta=label_meta,
      model_params=model_params_constructed,
      validation_metrics="Accuracy".split(","),
      save="model_save",
      batch_size=1,
      epoch=1,
      validation_steps=1,
      verbose=0,
      max_steps=train_max_steps,
      validation_start_delay_secs=0,
      validation_throttle_secs=0,
      save_checkpoints_steps=100,
      log_every_n_iter=10,
      is_pai="false" == "true",
      pai_table="",
      pai_val_table="")

==========Output==========
Start training using estimator model...
Traceback (most recent call last):
  File "<stdin>", line 136, in <module>
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train.py", line 126, in train
    save_checkpoints_steps, validation_metrics)
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 65, in estimator_train_and_save
    eval_start_delay_secs, eval_throttle_secs)
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 105, in estimator_train_compiled
    estimator.train(lambda: train_dataset_fn(), max_steps=train_max_steps)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 370, in train
    loss = self._train_model(input_fn, hooks, saving_listeners)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1160, in _train_model
    return self._train_model_default(input_fn, hooks, saving_listeners)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1190, in _train_model_default
    features, labels, ModeKeys.TRAIN, self.config)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/estimator.py", line 1148, in _call_model_fn
    model_fn_results = self._model_fn(features=features, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 764, in _model_fn
    batch_norm=batch_norm)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 573, in dnn_model_fn_v2
    mode=mode)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 508, in _dnn_model_fn_builder_v2
    name='dnn')
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_estimator/python/estimator/canned/dnn.py", line 328, in __init__
    name=hidden_shared_name)
  File "/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/keras/layers/core.py", line 995, in __init__
    self.units = int(units)
ValueError: invalid literal for int() with base 10: 'x'

What we want:

error: invalid attribute type: hidden_units of DNNClassifier should be an integer list like [30, 20, 10]

Required Argument Missed

SELECT * from iris.train TO TRAIN DNNClassifier WITH model.n_classes=3 LABEL class INTO iris.xgb;

What we have is

Traceback (most recent call last):

  File "<stdin>", line 136, in <module>

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train.py", line 126, in train

    save_checkpoints_steps, validation_metrics)

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 54, in estimator_train_and_save

    classifier = estimator(**model_params)

TypeError: __init__() missing 1 required positional argument: 'hidden_units'

ERROR: rpc error: code = Unknown desc = runSQLProgram error: failed: exit status 1
==========Program==========

import copy
import tensorflow as tf
from sqlflow_submitter.tensorflow.train import train
from sqlflow_submitter.tensorflow.get_tf_version import tf_is_version2
from tensorflow.estimator import (DNNClassifier,
                                  DNNRegressor,
                                  LinearClassifier,
                                  LinearRegressor,
                                  BoostedTreesClassifier,
                                  BoostedTreesRegressor,
                                  DNNLinearCombinedClassifier,
                                  DNNLinearCombinedRegressor)
if tf_is_version2():
    from tensorflow.keras.optimizers import Adadelta, Adagrad, Adam, Adamax, Ftrl, Nadam, RMSprop, SGD
    from tensorflow.keras.losses import BinaryCrossentropy, CategoricalCrossentropy, CategoricalHinge, CosineSimilarity, Hinge, Huber, KLDivergence, LogCosh, MeanAbsoluteError, MeanAbsolutePercentageError, MeanSquaredError, MeanSquaredLogarithmicError, Poisson, SparseCategoricalCrossentropy, SquaredHinge
else:
    from tensorflow.train import AdadeltaOptimizer, AdagradOptimizer, AdamOptimizer, FtrlOptimizer, RMSPropOptimizer, GradientDescentOptimizer, MomentumOptimizer
    from tensorflow.keras.losses import BinaryCrossentropy, CategoricalCrossentropy, CategoricalHinge, CosineSimilarity, Hinge, Huber, KLDivergence, LogCosh, MeanAbsoluteError, MeanAbsolutePercentageError, MeanSquaredError, MeanSquaredLogarithmicError, Poisson, SparseCategoricalCrossentropy, SquaredHinge
try:
    import sqlflow_models
except Exception as e:
    print("failed to import sqlflow_models: %s", e)

feature_column_names = [
"sepal_length",

"sepal_width",

"petal_length",

"petal_width",
]

# feature_column_names_map is used to determine the order of feature columns of each target:
# e.g. when using DNNLinearCombinedClassifer.
# feature_column_names_map will be saved to a single file when using PAI.
feature_column_names_map = dict()

feature_column_names_map["feature_columns"] = ["sepal_length","sepal_width","petal_length","petal_width",]

feature_metas = dict()

feature_metas["sepal_length"] = {
    "feature_name": "sepal_length",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["sepal_width"] = {
    "feature_name": "sepal_width",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["petal_length"] = {
    "feature_name": "petal_length",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

feature_metas["petal_width"] = {
    "feature_name": "petal_width",
    "dtype": "float32",
    "delimiter": "",
    "shape": [1],
    "is_sparse": "false" == "true"
}

label_meta = {
    "feature_name": "class",
    "dtype": "int64",
    "delimiter": "",
    "shape": [],
    "is_sparse": "false" == "true"
}

model_params=dict()

model_params["n_classes"]=3

# Construct optimizer objects to pass to model initializer.
# The original model_params is serializable (do not have tf.xxx objects).
model_params_constructed = copy.deepcopy(model_params)
for optimizer_arg in ["optimizer", "dnn_optimizer", "linear_optimizer"]:
    if optimizer_arg in model_params_constructed:
        model_params_constructed[optimizer_arg] = eval(model_params_constructed[optimizer_arg])

if "loss" in model_params_constructed:
    model_params_constructed["loss"] = eval(model_params_constructed["loss"])

# feature_columns_code will be used to save the training informations together
# with the saved model.
feature_columns_code = """{"feature_columns": [tf.feature_column.numeric_column("sepal_length", shape=[1]),
tf.feature_column.numeric_column("sepal_width", shape=[1]),
tf.feature_column.numeric_column("petal_length", shape=[1]),
tf.feature_column.numeric_column("petal_width", shape=[1])]}"""
feature_columns = eval(feature_columns_code)

train_max_steps = 0
train_max_steps = None if train_max_steps == 0 else train_max_steps

train(datasource="mysql://root:root@tcp(127.0.0.1:3306)/?maxAllowedPacket=0",
      estimator=DNNClassifier,
      select="""SELECT * from iris.train """,
      validation_select="""""",
      feature_columns=feature_columns,
      feature_column_names=feature_column_names,
      feature_metas=feature_metas,
      label_meta=label_meta,
      model_params=model_params_constructed,
      validation_metrics="Accuracy".split(","),
      save="model_save",
      batch_size=1,
      epoch=1,
      validation_steps=1,
      verbose=0,
      max_steps=train_max_steps,
      validation_start_delay_secs=0,
      validation_throttle_secs=0,
      save_checkpoints_steps=100,
      log_every_n_iter=10,
      is_pai="false" == "true",
      pai_table="",
      pai_val_table="")

==========Output==========
Start training using estimator model...
Traceback (most recent call last):
  File "<stdin>", line 136, in <module>
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train.py", line 126, in train
    save_checkpoints_steps, validation_metrics)
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/tensorflow/train_estimator.py", line 54, in estimator_train_and_save
    classifier = estimator(**model_params)
TypeError: __init__() missing 1 required positional argument: 'hidden_units'

What we want is

error: required argument missed: hidden_units of DNNClassifier must be provided

Column doesn't exist

SELECT * from iris.train TO TRAIN DNNClassifier WITH model.n_classes=3 LABEL asdf INTO iris.xgb ;
SELECT asdf from iris.train TO TRAIN DNNClassifier WITH model.n_classes=3 LABEL asdf INTO iris.xgb;

What we have

ERROR: rpc error: code = Unknown desc = runSQLProgram error: deriveLabel: LABEL COLUMN 'asdf' not found

ERROR: rpc error: code = Unknown desc = runSQLProgram error: Error 1054: Unknown column 'asdf' in 'field list'

What we want

error: unknown column: 'asdf' in 'field list' not found in 'iris.train'

error: unknown column: 'asdf' in 'label list' not found in 'iris.train'

inconsistent model type and statement

SELECT * from iris.train TO TRAIN xgboost.gbtree WITH objective=multi:softmax INTO iris.xgb;

What we have

Traceback (most recent call last):

  File "<stdin>", line 35, in <module>

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/train.py", line 53, in train

    for per_batch_dmatrix in dtrain:

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/dataset.py", line 60, in xgb_dataset

    feature_specs, label_spec)

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/dataset.py", line 96, in dump_dmatrix

    row_data = [str(item[1])] + row_data

IndexError: tuple index out of range

ERROR: rpc error: code = Unknown desc = runSQLProgram error: failed: exit status 1
==========Program==========

from sqlflow_submitter.xgboost.train import train
import json

model_params = json.loads('''{"booster":"gbtree","eta":0.3,"objective":"multi:softmax","validation.select":""}''')
train_params = json.loads('''{"num_boost_round":10}''')
feature_metas = json.loads('''{"class":{"feature_name":"class","dtype":"int64","delimiter":"","shape":[1],"is_sparse":false},"petal_length":{"feature_name":"petal_length","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"petal_width":{"feature_name":"petal_width","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"sepal_length":{"feature_name":"sepal_length","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"sepal_width":{"feature_name":"sepal_width","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false}}''')
label_meta = json.loads('''{"feature_name":"","dtype":"int64","delimiter":"","shape":null,"is_sparse":false}''')

feature_column_names = [
"sepal_length",

"sepal_width",

"petal_length",

"petal_width",

"class",
]

train(datasource='''mysql://root:root@tcp(127.0.0.1:3306)/?maxAllowedPacket=0''',
      select='''SELECT * from iris.train ''',
      model_params=model_params,
      train_params=train_params,
      feature_metas=feature_metas,
      feature_column_names=feature_column_names,
      label_meta=label_meta,
      validation_select='''''',
      disk_cache="false" == "true",
      batch_size=-1,
      epoch=1,
      is_pai="false" == "true",
      pai_train_table="",
      pai_validate_table="")

==========Output==========
Start training XGBoost model...
Traceback (most recent call last):
  File "<stdin>", line 35, in <module>
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/train.py", line 53, in train
    for per_batch_dmatrix in dtrain:
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/dataset.py", line 60, in xgb_dataset
    feature_specs, label_spec)
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/dataset.py", line 96, in dump_dmatrix
    row_data = [str(item[1])] + row_data
IndexError: tuple index out of range

What we want

error: 'xgboost.gbtree' is a supervised model, but LABEL clause is not provided.

Argument range checking

SELECT * from iris.train TO TRAIN xgboost.gbtree WITH objective=multi:softmax LABEL class INTO iris.xgb;

What we have:

Traceback (most recent call last):

  File "<stdin>", line 33, in <module>

  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/train.py", line 64, in train

    **train_params)

  File "/usr/local/lib/python3.6/dist-packages/xgboost/training.py", line 216, in train

    xgb_model=xgb_model, callbacks=callbacks)

  File "/usr/local/lib/python3.6/dist-packages/xgboost/training.py", line 74, in _train_internal

    bst.update(dtrain, i, obj)

  File "/usr/local/lib/python3.6/dist-packages/xgboost/core.py", line 1109, in update

    dtrain.handle))

  File "/usr/local/lib/python3.6/dist-packages/xgboost/core.py", line 176, in _check_call

    raise XGBoostError(py_str(_LIB.XGBGetLastError()))

xgboost.core.XGBoostError: value 0 for Parameter num_class should be greater equal to 1

ERROR: rpc error: code = Unknown desc = runSQLProgram error: failed: exit status 1
==========Program==========

from sqlflow_submitter.xgboost.train import train
import json

model_params = json.loads('''{"booster":"gbtree","eta":0.3,"objective":"multi:softmax","validation.select":""}''')
train_params = json.loads('''{"num_boost_round":10}''')
feature_metas = json.loads('''{"petal_length":{"feature_name":"petal_length","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"petal_width":{"feature_name":"petal_width","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"sepal_length":{"feature_name":"sepal_length","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false},"sepal_width":{"feature_name":"sepal_width","dtype":"float32","delimiter":"","shape":[1],"is_sparse":false}}''')
label_meta = json.loads('''{"feature_name":"class","dtype":"int64","delimiter":"","shape":[],"is_sparse":false}''')

feature_column_names = [
"sepal_length",

"sepal_width",

"petal_length",

"petal_width",
]

train(datasource='''mysql://root:root@tcp(127.0.0.1:3306)/?maxAllowedPacket=0''',
      select='''SELECT * from iris.train ''',
      model_params=model_params,
      train_params=train_params,
      feature_metas=feature_metas,
      feature_column_names=feature_column_names,
      label_meta=label_meta,
      validation_select='''''',
      disk_cache="false" == "true",
      batch_size=-1,
      epoch=1,
      is_pai="false" == "true",
      pai_train_table="",
      pai_validate_table="")

==========Output==========
Start training XGBoost model...
[12:39:48] 110x4 matrix with 440 entries loaded from train.txt_0
Traceback (most recent call last):
  File "<stdin>", line 33, in <module>
  File "/root/go/src/sqlflow.org/sqlflow/python/sqlflow_submitter/xgboost/train.py", line 64, in train
    **train_params)
  File "/usr/local/lib/python3.6/dist-packages/xgboost/training.py", line 216, in train
    xgb_model=xgb_model, callbacks=callbacks)
  File "/usr/local/lib/python3.6/dist-packages/xgboost/training.py", line 74, in _train_internal
    bst.update(dtrain, i, obj)
  File "/usr/local/lib/python3.6/dist-packages/xgboost/core.py", line 1109, in update
    dtrain.handle))
  File "/usr/local/lib/python3.6/dist-packages/xgboost/core.py", line 176, in _check_call
    raise XGBoostError(py_str(_LIB.XGBGetLastError()))
xgboost.core.XGBoostError: value 0 for Parameter num_class should be greater equal to 1

What we want:

error: invalid argument: num_class of xgboost.gbtree should be greater or equal to 1
Yancey1989 commented 4 years ago

Came from @llxxxll

Column Check Error

SELECT * from fake.train TO TRAIN xgboost.gbtree 
WITH objective=multi:softmax 
INTO fake.xgb;

XGboost's feature type should be numerical if some column type of fake.train is string, .e.g. cat, dog, we should tell users it does work on string feature column type.

What we have: TBD

What we want:

column type error: data type of input feature column `f` should be numerical for xgboost.gbtree estimator.
wangkuiyi commented 4 years ago

@Yancey1989 and @shendiaomo , thank you so much for the summarization of these error reporting. We definitely need to do something to make it better. Here follow my few cents.

Error Messaging in Go Code

In my mind, I have the following error handling schema in Go code.

  1. No panic and recover unless justifiable; instead,
  2. make every function return and error.

Given these two points, we can have error handling like shown in the following code snippet.

func a_func(params Something) (SomethingElse, error) {
   if e := do_some_thing(); e != nil {
      return nil, fmt.Errorf("a_func: waited to do .... but cannot do some thing:\n%v", e)
  }
  return result, nil
}

func grpc_server_Run() (Response, error) {
  r, e := a_func()
  if e != nil {
    return nil, fmt.Errorf("grpc_server_Run: wanted to ... but ...:\n%v", e)
   }
}

func grpc_client_main() {
  r, e := stub.Call(a_func)
  if e != nil {
    log.Error(e)
  }
}