Why null value in the table getting error while using lead function - sql

Question

I am getting this error - Error converting data type nvarchar to numeric.
I have data coming from a table and I need only two values from the table where I filter only the number (no alphanumeric so used the isnumeric(covrg_cd)=1). The input data looks like the first picture. The Row 1 will always be null and in other other rows, there may or may not be data. However, because row 1 is always null, the lead function is throwing this error: Error converting data type nvarchar to numeric, but the rate column is always in nvarchar. I am using LEAD function in SQL to get the paybandfrom & paybandto using the Rate from Input table and using row_number() to get the tier value.
Input table
out put must be like this..
I have my query like this
SELECT a.payband , a.[from] as pybdnfrom, (RIGHT('00000000000000000000' + CAST(A.[TO] AS VARCHAR),20)) AS pybndto , a.tier
FROM (SELECT DISTINCT A.RATE as payband, A.RATE as [from], CASE WHEN TIER <> 4 THEN A.[TO] ELSE 100000000.000 END AS [to], ROW_NUMBER() OVER(ORDER BY RATE) AS TIER
FROM(SELECT DISTINCT A.RATE, LEAD(SUM((CONVERT(NUMERIC(20,3), (A.RATE)))-0.010)) OVER(ORDER BY A.RATE) AS [TO], ROW_NUMBER() OVER(ORDER BY A.RATE) AS TIER
FROM (SELECT DISTINCT BN_RATE_KEY02 as RATE, COVRG_CD AS COVERAGE
from #tmppsRateCost
WHERE ISNUMERIC(COVRG_CD) = 1 AND COVRG_CD = '1')A GROUP BY A.RATE)A)A
ORDER BY 1
Any help would be appreciated.

Answer 1

The error is because the '' cannot be parsed as a number. It's not related to the LEAD.
If you want to keep that approach you can modify your query in this way (I just commented the parts I replaced):
SELECT a.payband
,a.[from] AS pybdnfrom
--,(RIGHT('00000000000000000000' + CAST(A.[TO] AS VARCHAR), 20)) AS pybndto
,CASE WHEN payband = '' THEN '' ELSE (RIGHT('00000000000000000000' + CAST(A.[TO] AS VARCHAR), 20)) END AS pybndto
,a.tier
FROM (
SELECT DISTINCT A.RATE AS payband
,A.RATE AS [from]
,CASE
WHEN TIER <> 5
THEN A.[TO]
ELSE 100000000.000
END AS [to]
,ROW_NUMBER() OVER (
ORDER BY RATE
) AS TIER
FROM (
SELECT DISTINCT A.RATE
--,LEAD(SUM((CONVERT(NUMERIC(20, 3), (A.RATE))) - 0.010)) OVER (
,LEAD(SUM((CONVERT(NUMERIC(20, 3), (CASE WHEN A.RATE = '' THEN '0.010' ELSE A.RATE END))) - 0.010)) OVER (
ORDER BY A.RATE
) AS [TO]
,ROW_NUMBER() OVER (
ORDER BY A.RATE
) AS TIER
FROM (
SELECT DISTINCT BN_RATE_KEY02 AS RATE
,COVRG_CD AS COVERAGE
FROM #tmppsRateCost
WHERE ISNUMERIC(COVRG_CD) = 1
AND COVRG_CD = '1'
) A
GROUP BY A.RATE
) A
) A
ORDER BY 1
Anyway I guess you might have a cleaner approach just by removing the empty line in the initial table.

Answer 2

Get rid of ISNUMERIC() and use TRY_CONVERT() insert of CONVERT(). In this condition:
WHERE ISNUMERIC(COVRG_CD) = 1 AND COVRG_CD = '1'
The ISNUMERIC() is just unneeded because you have an exact string comparison.
SELECT a.payband , a.[from] as pybdnfrom,
(RIGHT('00000000000000000000' + CAST(A.[TO] AS VARCHAR),20)) AS pybndto ,
a.tier
FROM (SELECT DISTINCT A.RATE as payband, A.RATE as [from],
CASE WHEN TIER <> 4 THEN A.[TO] ELSE 100000000.000 END AS [to],
ROW_NUMBER() OVER (ORDER BY RATE) AS TIER
FROM (SELECT DISTINCT A.RATE,
LEAD(SUM((TRY_CONVERT(NUMERIC(20,3), (A.RATE)))-0.010)) OVER (ORDER BY A.RATE) AS [TO],
ROW_NUMBER() OVER (ORDER BY A.RATE) AS TIER
FROM (SELECT DISTINCT BN_RATE_KEY02 as RATE, COVRG_CD AS COVERAGE
FROM #tmppsRateCost
WHERE COVRG_CD = '1'
)A
GROUP BY A.RATE
) A
) A
ORDER BY 1;