Getting "error_code": 1207 while validating before loading Redshift database

@amardeep2021 yes! well, at least give this a try …

Hi @amardeep2021

You still have this issue, but I now can’t reproduce the problem against a real or double field! I may have misspoken and have since tore down the pipeline that I’m sure was getting a 1207 validation error for a number with a trailing space trying to insert into a real field. Let’s back up.

I’ve attached a pipeline with just two snaps, a JSON Generate and Redshift BulkLoad.

Here’s the payload for the generator; notice trailing space in number value.

[
    {
        "id": 1,
        "bigdecimal": "595546349.04 "
    }
]

Let’s create the Redshift table to bulk load into. Observe number data type is an integer, so we expect problems.

CREATE TABLE t1207 (
    id         int,
    bigdecimal integer
);

Sure enough, the BulkLoad snap reports this error. We expects this.

[
  {
    "error": "Invalid digit, Value '.', Pos 9, Type: Integer",
    "status_code": "error",
    "line_number": 1,
    "column_name": "bigdecimal",
    "error_code": 1207,
    "column_value": "595546349.04",
    ...
  }
]

The type in the error message, Integer, indicates the field we’re trying to insert into. Ok, let’s drop table and re-create with real data type.

CREATE TABLE t1207 (
    id         int,
    bigdecimal real
);

Re-running the pipeline succeeds without having to trim the number value, which might make better sense, because you’re still having a problem after doing this indicating problem is something else.

[
  {
    "id": 1,
    "bigdecimal": 5.9554637E+8
  }
]

However, we have lost some precision. This is a big number. So, let’s drop table and re-create as double precision which is probably what we want.

CREATE TABLE t1207 (
    id         int,
    bigdecimal double precision
);

Pipeline succeeds and full precision is preserved.

[
  {
    "id": 1,
    "bigdecimal": 595546349.04
  }
]

I tried changing JDBC drivers from Postgres to Redshift and get same result with either. I thought maybe Redshift driver was handling space trimming, but apparently this isn’t a problem with either.

As I mentioned, BulkLoad first writes data to a CSV file, zips it, and then stores this in S3 for Redshift to source from. My Redshift account and BulkLoad snap don’t have any special properties defined.

Maybe, you can share a bit more about your configuration, settings, or anything else suspect.

redshift-bulkload-1207-error_2021_10_27.slp (5.6 KB)

What is the data type of annual_revenue? The error suggests it is an integer, and not a decimal number. Please verify for me.

@amardeep2021

The error you are receiving is from the Redshift database trying to insert a decimal into an integer field. Can you tell me the data type of the annual_revenue field ? It seems it is defined as an integer, but we’re trying to store a decimal number into field. So all the snap data transformation, etc, are relevant, but what really matters is the final value getting inserted into the target database, and it seems there is a mismatch here.

is annual_revenue an integer or a decimal? BIGINT will likely raise the same 1207 error trying to write a decimal value. It if should be an integer, then we should modify the pipeline to write this value as a integer.

@amardeep2021

DOUBLE PRECISION if we care about decimals. I see only two values (.84 and .14). If those can be stored as integers, then BIGINT. If BIGINT, we need to handle the decimal values in the pipeline by either rounding or truncating decimal values.

Oh, I also notice 0.00 and the .00 may be significant. DOUBLE PRECISION would be able to store all of these representations.

Hi @amardeep2021

I’m able to insert all of these values into a DOUBLE PRECISION field. When you say you tried double precision and it gave you an error, did you alter the data type of the database field to a DOUBLE PRECISION ? That’s what I thought you meant earlier, but now not sure we are referring to the same thing. The 1207 error you showed earlier was pretty specific about the reason for failure reporting the value and data type of the field it was trying to insert into.

CREATE TABLE t1207 (
    id         int,
    bigdecimal double precision
)

JSON payload.

[
    {
        "id": 1,
        "bigdecimal": "256800000000"
    },
    {
        "id": 2,
        "bigdecimal": 0.00
    },
    {
        "id": 3,
        "bigdecimal": "64290705800.14"
    },
    {
        "id": 4,
        "bigdecimal": "6515871900000"
    },
    {
        "id": 5,
        "bigdecimal": "76476000000"
    },
    {
        "id": 6,
        "bigdecimal": "87351502445.84"
    },
    {
        "id": 7,
        "bigdecimal": null 
    }
]

Be aware that you can also create a support ticket for help resolving your issue:

That said, I have no problem continuing to try to help here. I think you’re close.