Errors, Warnings, Quality Checks

Errors, Warnings, Quality Checks
Message Text
Message type
Bricks
Reason
How to solve
ERROR
All
This error triggers when you provide less inputs than expected. The most common situation is that no input connected at all, but also some brick may require multiple inputs with some configurations
Connect the needed number of inputs to the brick
ERROR
All
Most bricks require columns to be selected for processing. If none are provided for the selection the error is triggered. Also, some bricks may support multiple option selection by pressing “+” sign in the corresponding section.
Fill all options with columns. Also, you can press “-” sign near drop-down selection if you don’t need that many columns to be processed
ERROR
Filter Rows
Unexpected error happened while filtering with “Filter Rows” brick. Likely, unsupported values where provided for the selected data type. For example, such error might be triggered when you try to filter rows like this: 1 > “hello”
Check brick’s configuration, especially columns’ types. If needed, use “Typization” brick to convert columns to the required data type
WARNING
All
Brick produced empty data frame as a result. Also the empty data frame could have been provided as an input.
If it is expected - no treatment is needed, pipeline will continue running. But if you expected some results - you should check brick’s configuration if you misstyped something
ERROR
Export Data
You haven’t selected the SQL data source
You should select data source for data to be extracted
ERROR
Export Data
You haven’t selected the data table where you want the data to be extracted
Specify the data table you want the data to be extracted into
ERROR
Export Data
Such error are triggered when you selected “Fail” option in the “Export if the destination exists” setting and specified already existing data table
Either choose a different table name or select another option, like replace or append
ERROR
All
It means that you haven’t specified mandatory option/argument in the brick settings
Fill the option/argument with appropriate values
ERROR
Aggregate Data
You have selected an unsupported aggregation function for the column’s data type
Select a different function or use “Typization” brick to convert column to a supported data type
NOTIFICATION
All
A brick has rewritten already existing column
If it is expected - no treatment is needed, pipeline will continue running. Otherwise - rename resulting column to a different name
WARNING
Apply Function
“Apply Function” brick raises such warning if non-datetime like inputs or wrong date format were provided for any date- or time-related function
If expected - no treatment is needed, pipeline will continue running. Otherwise it is better to check the input columns.
ERROR
Apply Function
Such error occurs if non-numerical columns were provided
Check brick’s configuration or use “Typization” brick before that.
ERROR
Math Formula
Provided formula has incorrect syntax. Most common situation: wrong number of closing “)” or orthographic errors
Check the formula and fix the syntax.
ERROR
Math Formula
Provided formula contains wrong column names. Column names are specified by “{ }” brackets
Fix columns. For simplicity, you can see available columns on the left toolbar, also you can click on them to insert them into the formula
ERROR
Math Formula
Columns with an incorrect data type were provided to the formula’s functions
Check data types and use “Typization” brick if needed.
ERROR
Math Formula
Columns contain missing values. Math Formula applies functions by each row and because of that doesn’t correctly support missing values
Use “Missing Values Treatment” before the brick
NOTIFICATION
Missing Values Treatment
Regularly, such notification occurs when you run this brick for the first time.
It is not an issue. If needed, you can open the dashboard and manually select rules if you don’t agree with automatic suggestions
ERROR
Missing Values Treatment, Auto Data Preparation
You provided a fill value with a wrong data type
Use a different value or use “Typization” brick before the MVT
ERROR
Missing Values Treatment, Auto Data Preparation
You have selected filling function with unsupported data type
Use a different method or use “Typization” brick before the MVT
NOTIFICATION
Auto Data Preparation
A regular message to notify you that new auto-suggestion were applied
No actions are needed. You can check dashboard to apply own transformation rules
NOTIFICATION
Join Data
A message to notify you that both input data tables contain identical non-key columns and that they were renamed to eliminate confusion
If expected - no additional actions are needed. Potentially, you either should rename columns, use them as keys or filter them out if needed.
NOTIFICATION
Join Data
Sometimes key columns provided to “Join Data” brick as inputs may be of different but compatible data types (for example, category and string). This message notifies your about this and also that brick automatically converted to a specific data type
No additional actions are needed. Optionally, you can manually convert data types using the “Typization” brick
ERROR
Parse JSON
Parse JSON support only one row tables as the input
You should provide single rows only, no other options are supported
ERROR
Parse JSON
Wrong data format was provided to the “Parse JSON” input
Check if your input data is in JSON-format and doesn’t contain any errors
NOTIFICATION
Shift Data
“Shift Data” saves data to a new column. This notification says how it is called
No actions are needed
ERROR
Typization
It means that some rows contain values that cannot be converted to the selected data type
Preprocess the data or choose a different data type
ERROR
Encoding
One-hot encoding is meant for categorical column with small number of distinct values. A larger number will cause issues with memory
Consider using other
ERROR
Split Data
If you want to stratify columns during the split, you should also shuffle the data
Enable the “Shuffle” checkbox
ERROR
Split Data
You can’t split dataframes with one row
Check whether the data contains all rows and none are missed. It is not possible to split 1 element into two different dataframes
ERROR
All
Some bricks support only specific data types and not the others
Use other columns or use “Typization” brick to convert data types to supported
ERROR
Split Data
You can’t split dataframes with one row
Check whether the data contains all rows and none are missed. It is not possible to split 1 element into two different dataframes
ERROR
Split Data
If you want to stratify columns during the split, you should also shuffle the data
Enable the “Shuffle” checkbox
ERROR
Compare Data
You have provided columns/values with incompatible data types. For example, you might tried to compare string with a float
Use “Typization” brick before this
ERROR
Iterative Calculations
This brick doesn’t support non-numerical columns
Use a different column or apply “Typization” brick beforehand
NOTIFICATION
All
Notifies you that new column will be created after the brick run, also specifies how will it be called
No further actions are needed
ERROR
Rolling Function
This brick doesn’t support non-numerical columns
Use a different column or apply “Typization” brick beforehand
ERROR
Encoding
One-hot encoding is meant for categorical columns with limited number of unique values and won’t properly function with a large number of them
Consider using other types of encodings
ERROR
Encoding
Binary encoding doesn’t support missing values because it is impossible to know beforehand which value to assign to them
Consider performing missing values treatment or use another encoding method
ERROR
Encoding
Binary encoding support only two distinct values
Use another encoding method, namely - label encoding
ERROR
Parse Datetime
Provided column has inappropriate date format
You should manually preprocess provided column to conventional date formats
ERROR
All
Some bricks support only specific data types and not the others
Use other columns or use “Typization” brick to convert data types to supported
ERROR
Typization
It means that some rows contain nonconvertable values to the selected data type
Preprocess values before applying typization
ERROR
Encoding
One-hot encoding is meant for categorical columns with low number of unique values. Larger number will cause extreme performance and memory issues
Consider using other types of encoding
ERROR
Encoding
Binary encoding supports only columns with two unique values
Use a different type of encoding, namely label encoding
ERROR
Encoding
Binary encoding does not support missing values because it is hard to predict to which class NaNs should be assigned
Use other types of encoding or use “Missing Values Treatment” beforehand
ERROR
Some models
Not all models support string values
Remove string columns or preprocess them using encoding or typization to convert them to numeric data type
ERROR
Some models
Not all models support missing values
Preprocess data by removing missing values with “Missing Values Treatment” brick
ERROR
For Loop
Input either is not connected or selected in the settings
Connect and/or select input in the settings
ERROR
For Loop
Input is connected but no columns to iterate are selected
Select at least one column to iterate
ERROR
For Loop
For-loop inside for-loop is not supported yet due computational costs
Remove for loop
ERROR
For Loop
Pipeline inside for-loop is not supported yet due to computational costs
Remove the pipeline brick from the for loop (you can unwrap pipelines to extract bricks inside)