[BUG]: Error in density.default(BetaIV.in, weights = weights, bw = h) : non-finite 'bw'

[Julie1130]

When I run the mr(datm), it gave me this error.

Error in density.default(BetaIV.in, weights = weights, bw = h) : non-finite 'bw'

[Julie1130]

@explodecomputer Could you please take a look? Thanks!

[Julie1130]

@mathspete Could you please take a look? Thanks!

[remlapmot]

You have not provided a reproducible example. Please provide a full code example including the data which recreates the error otherwise it's very time consuming for us just to recreate your error.

(We ask for a reproducible example in the issue template - but you've deleted that.)

And please don't @ mention people without good reason.

[Julie1130]

I am really sorry. Here are the codes:
data1 = read_csv("D:\Julie\Data\smaller_p_physical.csv")
exposure_dat = format_data(data1, type = "exposure",
snp_col = "SNP rsid", chr_col = "Chr",
beta_col = "beta", se_col = "SE", pval_col = "P",
effect_allele_col = "EA", other_allele_col = "OA",
eaf_col = "EAF",ncase_col="n",pos_col = "Position")
exposure_dat<-clump_data(exposure_dat)
exposure_dat$r2<-2 * exposure_dat$eaf.exposure * (1-exposure_dat$eaf.exposure) * exposure_dat$beta.exposure * exposure_dat$beta.exposure
exposure_dat$r2
sum(exposure_dat$r2)
outcome_datm_male <- extract_outcome_data(exposure_dat$SNP, c('ebi-a-GCST006701'), proxies = 1, rsq = 0.8, align_alleles = 1, palindromes = 1, maf_threshold = 0.45)

datm_male<- harmonise_data(exposure_dat, outcome_datm_male)
print(mr(datm_male)) #results from MR base

---

it works fine after harmonization, but mr(datm_male) gave the bug, is it because several of the beta.exposure is 0 in my data?

[Julie1130]

You have not provided a reproducible example. Please provide a full code example including the data which recreates the error otherwise it's very time consuming for us just to recreate your error.

(We ask for a reproducible example in the issue template - but you've deleted that.)

And please don't @ mention people without good reason.

I am really sorry. I have provided the codes above.

[Julie1130]

Additionally, when I use "MendelianRandomization" package, it works fine. So it's not because the number of SNPs is too small.

[remlapmot]

The example is not reproducible because we don't have your dataset smaller_p_physical.csv.

[Julie1130]

small_p_physical.csv
Here you are. Hopefully it can be downloaded. If not, please let me know. Thank you very much indeed!